Publications
Search
761.
Ghosh, Sayan; Laksana, Eugene; Morency, Louis-Philippe; Scherer, Stefan
Representation Learning for Speech Emotion Recognition Journal Article
In: Interspeech 2016, pp. 3603–3607, 2016.
@article{ghosh_representation_2016,
title = {Representation Learning for Speech Emotion Recognition},
author = {Sayan Ghosh and Eugene Laksana and Louis-Philippe Morency and Stefan Scherer},
url = {https://www.researchgate.net/publication/307889274_Representation_Learning_for_Speech_Emotion_Recognition},
doi = {10.21437},
year = {2016},
date = {2016-09-01},
journal = {Interspeech 2016},
pages = {3603–3607},
abstract = {Speech emotion recognition is an important problem with applications as varied as human-computer interfaces and affective computing. Previous approaches to emotion recognition have mostly focused on extraction of carefully engineered features and have trained simple classifiers for the emotion task. There has been limited effort at representation learning for affect recognition, where features are learnt directly from the signal waveform or spectrum. Prior work also does not investigate the effect of transfer learning from affective attributes such as valence and activation to categorical emotions. In this paper, we investigate emotion recognition from spectrogram features extracted from the speech and glottal flow signals; spectrogram encoding is performed by a stacked autoencoder and an RNN (Recurrent Neural Network) is used for classification of four primary emotions. We perform two experiments to improve RNN training : (1) Representation Learning - Model training on the glottal flow signal to investigate the effect of speaker and phonetic invariant features on classification performance (2) Transfer Learning - RNN training on valence and activation, which is adapted to a four emotion classification task. On the USC-IEMOCAP dataset, our proposed approach achieves a performance comparable to the state of the art speech emotion recognition systems.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Speech emotion recognition is an important problem with applications as varied as human-computer interfaces and affective computing. Previous approaches to emotion recognition have mostly focused on extraction of carefully engineered features and have trained simple classifiers for the emotion task. There has been limited effort at representation learning for affect recognition, where features are learnt directly from the signal waveform or spectrum. Prior work also does not investigate the effect of transfer learning from affective attributes such as valence and activation to categorical emotions. In this paper, we investigate emotion recognition from spectrogram features extracted from the speech and glottal flow signals; spectrogram encoding is performed by a stacked autoencoder and an RNN (Recurrent Neural Network) is used for classification of four primary emotions. We perform two experiments to improve RNN training : (1) Representation Learning - Model training on the glottal flow signal to investigate the effect of speaker and phonetic invariant features on classification performance (2) Transfer Learning - RNN training on valence and activation, which is adapted to a four emotion classification task. On the USC-IEMOCAP dataset, our proposed approach achieves a performance comparable to the state of the art speech emotion recognition systems.
762.
Ghosh, Sayan; Laksana, Eugene; Morency, Louis-Philippe; Scherer, Stefen
An Unsupervised Approach to Glottal Inverse Filtering Proceedings Article
In: Proceedings of the 2016 24th European Signal Processing Conference (EUSIPCO), Budapest, Hungary, 2016.
@inproceedings{ghosh_unsupervised_2016,
title = {An Unsupervised Approach to Glottal Inverse Filtering},
author = {Sayan Ghosh and Eugene Laksana and Louis-Philippe Morency and Stefen Scherer},
url = {http://www.eurasip.org/Proceedings/Eusipco/Eusipco2016/papers/1570252319.pdf},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the 2016 24th European Signal Processing Conference (EUSIPCO)},
address = {Budapest, Hungary},
abstract = {The extraction of the glottal volume velocity waveform from voiced speech is a well-known example of a sparse signal recovery problem. Prior approaches have mostly used wellengineered speech processing or convex L1-optimization methods to solve the inverse filtering problem. In this paper, we describe a novel approach to modeling the human vocal tract using an unsupervised dictionary learning framework. We make the assumption of an all-pole model of the vocal tract, and derive an L1 regularized least squares loss function for the all-pole approximation. To evaluate the quality of the extracted glottal volume velocity waveform, we conduct experiments on real-life speech datasets, which include vowels and multi-speaker phonetically balanced utterances. We find that the the unsupervised model learns meaningful dictionaries of vocal tracts, and the proposed data-driven unsupervised framework achieves a performance comparable to the IAIF (Iterative Adaptive Inverse Filtering) glottal flow extraction approach.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The extraction of the glottal volume velocity waveform from voiced speech is a well-known example of a sparse signal recovery problem. Prior approaches have mostly used wellengineered speech processing or convex L1-optimization methods to solve the inverse filtering problem. In this paper, we describe a novel approach to modeling the human vocal tract using an unsupervised dictionary learning framework. We make the assumption of an all-pole model of the vocal tract, and derive an L1 regularized least squares loss function for the all-pole approximation. To evaluate the quality of the extracted glottal volume velocity waveform, we conduct experiments on real-life speech datasets, which include vowels and multi-speaker phonetically balanced utterances. We find that the the unsupervised model learns meaningful dictionaries of vocal tracts, and the proposed data-driven unsupervised framework achieves a performance comparable to the IAIF (Iterative Adaptive Inverse Filtering) glottal flow extraction approach.
763.
Buckwalter, J. Galen; Castellani, Brian; Mcewen, Bruce; Karlamangla, Arun S.; Rizzo, Albert A.; John, Bruce; O'donnell, Kyle; Seeman, Teresa
Allostatic Load as a Complex Clinical Construct: A Case-Based Computational Modeling Approach Journal Article
In: Complexity, vol. 21, no. S1, pp. 291–306, 2016, ISSN: 10762787.
@article{galen_buckwalter_allostatic_2016,
title = {Allostatic Load as a Complex Clinical Construct: A Case-Based Computational Modeling Approach},
author = {J. Galen Buckwalter and Brian Castellani and Bruce Mcewen and Arun S. Karlamangla and Albert A. Rizzo and Bruce John and Kyle O'donnell and Teresa Seeman},
url = {http://doi.wiley.com/10.1002/cplx.21743},
doi = {10.1002/cplx.21743},
issn = {10762787},
year = {2016},
date = {2016-09-01},
journal = {Complexity},
volume = {21},
number = {S1},
pages = {291–306},
abstract = {Allostatic load (AL) is a complex clinical construct, providing a unique window into the cumulative impact of stress. However, due to its inherent complexity, AL presents two major measurement challenges to conventional statistical modeling (the field’s dominant methodology): it is comprised of a complex causal network of bioallostatic systems, represented by an even larger set of dynamic biomarkers; and, it is situated within a web of antecedent socioecological systems, linking AL to differences in health outcomes and disparities. To address these challenges, we employed casebased computational modeling (CBM), which allowed us to make four advances: (1) we developed a multisystem, 7-factor (20 biomarker) model of AL’s network of allostatic systems; (2) used it to create a catalog of nine different clinical AL profiles (causal pathways); (3) linked each clinical profile to a typology of 23 health outcomes; and (4) explored our results (post hoc) as a function of gender, a key socioecological factor. In terms of highlights, (a) the Healthy clinical profile had few health risks; (b) the pro-inflammatory profile linked to high blood pressure and diabetes; (c) Low Stress Hormones linked to heart disease, TIA/Stroke, diabetes, and circulation problems; and (d) high stress hormones linked to heart disease and high blood pressure. Post hoc analyses also found that males were overrepresented on the High Blood Pressure (61.2%), Metabolic Syndrome (63.2%), High Stress Hormones (66.4%), and High Blood Sugar (57.1%); while females were overrepresented on the Healthy (81.9%), Low Stress Hormones (66.3%), and Low Stress Antagonists (stress buffers) (95.4%) profiles.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Allostatic load (AL) is a complex clinical construct, providing a unique window into the cumulative impact of stress. However, due to its inherent complexity, AL presents two major measurement challenges to conventional statistical modeling (the field’s dominant methodology): it is comprised of a complex causal network of bioallostatic systems, represented by an even larger set of dynamic biomarkers; and, it is situated within a web of antecedent socioecological systems, linking AL to differences in health outcomes and disparities. To address these challenges, we employed casebased computational modeling (CBM), which allowed us to make four advances: (1) we developed a multisystem, 7-factor (20 biomarker) model of AL’s network of allostatic systems; (2) used it to create a catalog of nine different clinical AL profiles (causal pathways); (3) linked each clinical profile to a typology of 23 health outcomes; and (4) explored our results (post hoc) as a function of gender, a key socioecological factor. In terms of highlights, (a) the Healthy clinical profile had few health risks; (b) the pro-inflammatory profile linked to high blood pressure and diabetes; (c) Low Stress Hormones linked to heart disease, TIA/Stroke, diabetes, and circulation problems; and (d) high stress hormones linked to heart disease and high blood pressure. Post hoc analyses also found that males were overrepresented on the High Blood Pressure (61.2%), Metabolic Syndrome (63.2%), High Stress Hormones (66.4%), and High Blood Sugar (57.1%); while females were overrepresented on the Healthy (81.9%), Low Stress Hormones (66.3%), and Low Stress Antagonists (stress buffers) (95.4%) profiles.
764.
Ravi, Satheesh; Artstein, Ron
Language Portability for Dialogue Systems: Translating a Question-Answering System from English into Tamil Proceedings Article
In: Proceedings of the SIGDIAL 2016 Conference, pp. 111–116, Association for Computational Linguistics, Los Angeles, CA, 2016.
@inproceedings{ravi_language_2016,
title = {Language Portability for Dialogue Systems: Translating a Question-Answering System from English into Tamil},
author = {Satheesh Ravi and Ron Artstein},
url = {http://www.aclweb.org/anthology/W16-3614},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the SIGDIAL 2016 Conference},
pages = {111–116},
publisher = {Association for Computational Linguistics},
address = {Los Angeles, CA},
abstract = {A training and test set for a dialogue system in the form of linked questions and responses is translated from English into Tamil. Accuracy of identifying an appropriate response in Tamil is 79%, compared to the English accuracy of 89%, suggesting that translation can be useful to start up a dialogue system. Machine translation of Tamil inputs into English also results in 79% accuracy. However, machine translation of the English training data into Tamil results in a drop in accuracy to 54% when tested on manually authored Tamil, indicating that there is still a large gap before machine translated dialogue systems can interact with human users.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
A training and test set for a dialogue system in the form of linked questions and responses is translated from English into Tamil. Accuracy of identifying an appropriate response in Tamil is 79%, compared to the English accuracy of 89%, suggesting that translation can be useful to start up a dialogue system. Machine translation of Tamil inputs into English also results in 79% accuracy. However, machine translation of the English training data into Tamil results in a drop in accuracy to 54% when tested on manually authored Tamil, indicating that there is still a large gap before machine translated dialogue systems can interact with human users.
765.
Manuvinakurike, Ramesh; Paetzel, Maike; Qu, Cheng; Schlangen, David; DeVault, David
Toward incremental dialogue act segmentation in fast-paced interactive dialogue systems Proceedings Article
In: Proceedings of the 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pp. 252–262, Association for Computational Linguistics, Los Angeles, CA, 2016.
@inproceedings{manuvinakurike_toward_2016,
title = {Toward incremental dialogue act segmentation in fast-paced interactive dialogue systems},
author = {Ramesh Manuvinakurike and Maike Paetzel and Cheng Qu and David Schlangen and David DeVault},
url = {http://www.aclweb.org/anthology/W16-3632},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue},
pages = {252–262},
publisher = {Association for Computational Linguistics},
address = {Los Angeles, CA},
abstract = {In this paper, we present and evaluate an approach to incremental dialogue act (DA) segmentation and classification. Our approach utilizes prosodic, lexico-syntactic and contextual features, and achieves an encouraging level of performance in offline corpus-based evaluation as well as in simulated human-agent dialogues. Our approach uses a pipeline of sequential processing steps, and we investigate the contribution of different processing steps to DA segmentation errors. We present our results using both existing and new metrics for DA segmentation. The incremental DA segmentation capability described here may help future systems to allow more natural speech from users and enable more natural patterns of interaction.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we present and evaluate an approach to incremental dialogue act (DA) segmentation and classification. Our approach utilizes prosodic, lexico-syntactic and contextual features, and achieves an encouraging level of performance in offline corpus-based evaluation as well as in simulated human-agent dialogues. Our approach uses a pipeline of sequential processing steps, and we investigate the contribution of different processing steps to DA segmentation errors. We present our results using both existing and new metrics for DA segmentation. The incremental DA segmentation capability described here may help future systems to allow more natural speech from users and enable more natural patterns of interaction.
766.
Manuvinakurike, Ramesh; Kennington, Casey; DeVault, David; Schlangen, David
Real-Time Understanding of Complex Discriminative Scene Descriptions Proceedings Article
In: Proceedings of the 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pp. 232–241, Association for Computational Linguistics, Los Angeles, CA, 2016.
@inproceedings{manuvinakurike_real-time_2016,
title = {Real-Time Understanding of Complex Discriminative Scene Descriptions},
author = {Ramesh Manuvinakurike and Casey Kennington and David DeVault and David Schlangen},
url = {http://www.aclweb.org/anthology/W16-3630},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue},
pages = {232–241},
publisher = {Association for Computational Linguistics},
address = {Los Angeles, CA},
abstract = {Real-world scenes typically have complex structure, and utterances about them consequently do as well. We devise and evaluate a model that processes descriptions of complex configurations of geometric shapes and can identify the described scenes among a set of candidates, including similar distractors. The model works with raw images of scenes, and by design can work word-by-word incrementally. Hence, it can be used in highly-responsive interactive and situated settings. Using a corpus of descriptions from game-play between human subjects (who found this to be a challenging task), we show that reconstruction of description structure in our system contributes to task success and supports the performance of the word-based model of grounded semantics that we use.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Real-world scenes typically have complex structure, and utterances about them consequently do as well. We devise and evaluate a model that processes descriptions of complex configurations of geometric shapes and can identify the described scenes among a set of candidates, including similar distractors. The model works with raw images of scenes, and by design can work word-by-word incrementally. Hence, it can be used in highly-responsive interactive and situated settings. Using a corpus of descriptions from game-play between human subjects (who found this to be a challenging task), we show that reconstruction of description structure in our system contributes to task success and supports the performance of the word-based model of grounded semantics that we use.
767.
Konovalov, Vasily; Melamud, Oren; Artstein, Ron; Dagan, Ido
Collecting Better Training Data using Biased Agent Policies in Negotiation Dialogues Proceedings Article
In: Proceedings of WOCHAT, the Second Workshop on Chatbots and Conversational Agent Technologies, Zerotype, Los Angeles, 2016.
@inproceedings{konovalov_collecting_2016,
title = {Collecting Better Training Data using Biased Agent Policies in Negotiation Dialogues},
author = {Vasily Konovalov and Oren Melamud and Ron Artstein and Ido Dagan},
url = {http://workshop.colips.org/wochat/documents/RP-270.pdf},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of WOCHAT, the Second Workshop on Chatbots and Conversational Agent Technologies},
publisher = {Zerotype},
address = {Los Angeles},
abstract = {When naturally occurring data is characterized by a highly skewed class distribution, supervised learning often benefits from reducing this skew. Human-agent dialogue data is commonly highly skewed when using standard agent policies. Hence, we suggest that agent policies need to be reconsidered in the context of training data collection. Specifically, in this work we implemented biased agent policies that are optimized for data collection in the negotiation domain. Empirical evaluations show that our method is successful in collecting a reasonably balanced corpus in the highly skewed Job-Candidate domain. Furthermore, using this balanced corpus to train a negotiation intent classifier yields notable performance improvements relative to naturally distributed data.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
When naturally occurring data is characterized by a highly skewed class distribution, supervised learning often benefits from reducing this skew. Human-agent dialogue data is commonly highly skewed when using standard agent policies. Hence, we suggest that agent policies need to be reconsidered in the context of training data collection. Specifically, in this work we implemented biased agent policies that are optimized for data collection in the negotiation domain. Empirical evaluations show that our method is successful in collecting a reasonably balanced corpus in the highly skewed Job-Candidate domain. Furthermore, using this balanced corpus to train a negotiation intent classifier yields notable performance improvements relative to naturally distributed data.
768.
Gratch, Jonathan; DeVault, David; Lucas, Gale
The Benefits of Virtual Humans for Teaching Negotiation Proceedings Article
In: Proceedings of the 16th International Conference on Intelligent Virtual Agents (IVA), 2016, Springer, Los Angeles, CA, 2016.
@inproceedings{gratch_benefits_2016,
title = {The Benefits of Virtual Humans for Teaching Negotiation},
author = {Jonathan Gratch and David DeVault and Gale Lucas},
url = {http://iva2016.ict.usc.edu/wp-content/uploads/Papers/100110276.pdf},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the 16th International Conference on Intelligent Virtual Agents (IVA), 2016},
publisher = {Springer},
address = {Los Angeles, CA},
abstract = {This article examines the potential for teaching negotiation with virtual humans. Many people find negotiations to be aversive. We conjecture that stu-dents may be more comfortable practicing negotiation skills with an agent than with another person. We test this using the Conflict Resolution Agent, a semi-automated virtual human that negotiates with people via natural language. In a between-participants design, we independently manipulated two pedagogically-relevant factors while participants engaged in repeated negotiations with the agent: perceived agency (participants either believed they were negotiating with a computer program or another person) and pedagogical feedback (participants received instructional advice or no advice between negotiations). Findings indi-cate that novice negotiators were more comfortable negotiating with a computer program (they self-reported more comfort and punished their opponent less of-ten) and expended more effort on the exercise following instructional feedback (both in time spent and in self-reported effort). These findings lend support to the notion of using virtual humans to teach interpersonal skills.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This article examines the potential for teaching negotiation with virtual humans. Many people find negotiations to be aversive. We conjecture that stu-dents may be more comfortable practicing negotiation skills with an agent than with another person. We test this using the Conflict Resolution Agent, a semi-automated virtual human that negotiates with people via natural language. In a between-participants design, we independently manipulated two pedagogically-relevant factors while participants engaged in repeated negotiations with the agent: perceived agency (participants either believed they were negotiating with a computer program or another person) and pedagogical feedback (participants received instructional advice or no advice between negotiations). Findings indi-cate that novice negotiators were more comfortable negotiating with a computer program (they self-reported more comfort and punished their opponent less of-ten) and expended more effort on the exercise following instructional feedback (both in time spent and in self-reported effort). These findings lend support to the notion of using virtual humans to teach interpersonal skills.
769.
Rizzo, AA; Lucas, G; Gratch, J; Stratou, G; Morency, L-P; Shilling, R; Hartholt, A; Scherer, S
Clinical interviewing by a virtual human agent with automatic behavior analysis Proceedings Article
In: Proceedings of The 2016 Proceedings of the International Conference on Disability, Virtual Reality and Associated Technologies, pp. 57–64, ICDVRAT and the University of Reading, Los Angeles, CA, 2016, ISBN: 978-0-7049-1547-3.
@inproceedings{rizzo_clinical_2016,
title = {Clinical interviewing by a virtual human agent with automatic behavior analysis},
author = {AA Rizzo and G Lucas and J Gratch and G Stratou and L-P Morency and R Shilling and A Hartholt and S Scherer},
url = {http://centaur.reading.ac.uk/66645/8/ICDVRAT2016_Full_Proceedings_11th%20_Conf.pdf},
isbn = {978-0-7049-1547-3},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of The 2016 Proceedings of the International Conference on Disability, Virtual Reality and Associated Technologies},
pages = {57–64},
publisher = {ICDVRAT and the University of Reading},
address = {Los Angeles, CA},
abstract = {SimSensei is a Virtual Human (VH) interviewing platform that uses off-the-shelf sensors (i.e., webcams, Microsoft Kinect and a microphone) to capture and interpret real-time audiovisual behavioral signals from users interacting with the VH system. The system was specifically designed for clinical interviewing and health care support by providing a face-to-face interaction between a user and a VH that can automatically react to the inferred state of the user through analysis of behavioral signals gleaned from the user’s facial expressions, body gestures and vocal parameters. Akin to how non-verbal behavioral signals have an impact on human-to-human interaction and communication, SimSensei aims to capture and infer user state from signals generated from user non-verbal communication to improve engagement between a VH and a user and to quantify user state from the data captured across a 20 minute interview. As well, previous research with SimSensei indicates that users engaging with this automated system, have less fear of evaluation and self-disclose more personal information compare to when they believe the VH agent is actually an avatar being operated by a “wizard of oz” human-in-the-loop (Lucas et al., 2014). The current study presents results from a sample of military service members (SMs) who were interviewed within the SimSensei system before and after a deployment to Afghanistan. Results indicate that SMs reveal more PTSD symptoms to the SimSensei VH agent than they self-report on the Post Deployment Health Assessment. Pre/Post deployment facial expression analysis indicated more sad expressions and fewer happy expressions at post deployment.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
SimSensei is a Virtual Human (VH) interviewing platform that uses off-the-shelf sensors (i.e., webcams, Microsoft Kinect and a microphone) to capture and interpret real-time audiovisual behavioral signals from users interacting with the VH system. The system was specifically designed for clinical interviewing and health care support by providing a face-to-face interaction between a user and a VH that can automatically react to the inferred state of the user through analysis of behavioral signals gleaned from the user’s facial expressions, body gestures and vocal parameters. Akin to how non-verbal behavioral signals have an impact on human-to-human interaction and communication, SimSensei aims to capture and infer user state from signals generated from user non-verbal communication to improve engagement between a VH and a user and to quantify user state from the data captured across a 20 minute interview. As well, previous research with SimSensei indicates that users engaging with this automated system, have less fear of evaluation and self-disclose more personal information compare to when they believe the VH agent is actually an avatar being operated by a “wizard of oz” human-in-the-loop (Lucas et al., 2014). The current study presents results from a sample of military service members (SMs) who were interviewed within the SimSensei system before and after a deployment to Afghanistan. Results indicate that SMs reveal more PTSD symptoms to the SimSensei VH agent than they self-report on the Post Deployment Health Assessment. Pre/Post deployment facial expression analysis indicated more sad expressions and fewer happy expressions at post deployment.
770.
Marge, Matthew; Bonial, Claire; Pollard, Kimberly A.; Artstein, Ron; Byrne, Brendan; Hill, Susan G.; Voss, Clare; Traum, David
Assessing Agreement in Human-Robot Dialogue Strategies: A Tale of TwoWizards Proceedings Article
In: Proceedings of The Sixteenth International Conference on Intelligent Virtual Agents (IVA 2016),, Springer, Los Angeles, CA, 2016.
@inproceedings{marge_assessing_2016,
title = {Assessing Agreement in Human-Robot Dialogue Strategies: A Tale of TwoWizards},
author = {Matthew Marge and Claire Bonial and Kimberly A. Pollard and Ron Artstein and Brendan Byrne and Susan G. Hill and Clare Voss and David Traum},
url = {http://iva2016.ict.usc.edu/wp-content/uploads/Papers/100110460.pdf},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of The Sixteenth International Conference on Intelligent Virtual Agents (IVA 2016),},
publisher = {Springer},
address = {Los Angeles, CA},
abstract = {The Wizard-of-Oz (WOz) method is a common experimental technique in virtual agent and human-robot dialogue research for eliciting natural communicative behavior from human partners when full autonomy is not yet possible. For the first phase of our research reported here, wizards play the role of dialogue manager, acting as a robot’s dialogue processing. We describe a novel step within WOz methodology that incorporates two wizards and control sessions: the wizards function much like corpus annotators, being asked to make independent judgments on how the robot should respond when receiving the same verbal commands in separate trials. We show that inter-wizard discussion after the control sessions and the resolution with a reconciled protocol for the follow-on pilot sessions successfully impacts wizard behaviors and significantly aligns their strategies. We conclude that, without control sessions, we would have been unlikely to achieve both the natural diversity of expression that comes with multiple wizards and a better protocol for modeling an automated system.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The Wizard-of-Oz (WOz) method is a common experimental technique in virtual agent and human-robot dialogue research for eliciting natural communicative behavior from human partners when full autonomy is not yet possible. For the first phase of our research reported here, wizards play the role of dialogue manager, acting as a robot’s dialogue processing. We describe a novel step within WOz methodology that incorporates two wizards and control sessions: the wizards function much like corpus annotators, being asked to make independent judgments on how the robot should respond when receiving the same verbal commands in separate trials. We show that inter-wizard discussion after the control sessions and the resolution with a reconciled protocol for the follow-on pilot sessions successfully impacts wizard behaviors and significantly aligns their strategies. We conclude that, without control sessions, we would have been unlikely to achieve both the natural diversity of expression that comes with multiple wizards and a better protocol for modeling an automated system.
771.
Chollet, Mathieu; Chandrashekhar, Nithin; Shapiro, Ari; Morency, Louis-Philippe; Scherer, Stefan
Manipulating the Perception of Virtual Audiences using Crowdsourced Behaviors Proceedings Article
In: Proceedings of the IVA 2016 : Intelligent Virtual Agents Conference, Springer, Los Angeles, CA, 2016.
@inproceedings{chollet_manipulating_2016,
title = {Manipulating the Perception of Virtual Audiences using Crowdsourced Behaviors},
author = {Mathieu Chollet and Nithin Chandrashekhar and Ari Shapiro and Louis-Philippe Morency and Stefan Scherer},
url = {http://iva2016.ict.usc.edu/wp-content/uploads/Papers/100110162.pdf},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the IVA 2016 : Intelligent Virtual Agents Conference},
publisher = {Springer},
address = {Los Angeles, CA},
abstract = {Virtual audiences are used for training public speaking and mitigating anxiety related to it. However, research has been scarce on studying how virtual audiences are perceived and which non-verbal behaviors should be used to make such an audience appear in particular states, such as boredom or engagement. Recently, crowdsourcing methods have been proposed for collecting data for building virtual agents' behavior models. In this paper, we use crowdsourcing for creating and evaluating a nonverbal behaviors generation model for virtual audiences. We show that our model successfully expresses relevant audience states (i.e. low to high arousal, negative to positive valence), and that the overall impression exhibited by the virtual audience can be controlled my manipulating the amount of individual audience members that display a congruent state.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Virtual audiences are used for training public speaking and mitigating anxiety related to it. However, research has been scarce on studying how virtual audiences are perceived and which non-verbal behaviors should be used to make such an audience appear in particular states, such as boredom or engagement. Recently, crowdsourcing methods have been proposed for collecting data for building virtual agents' behavior models. In this paper, we use crowdsourcing for creating and evaluating a nonverbal behaviors generation model for virtual audiences. We show that our model successfully expresses relevant audience states (i.e. low to high arousal, negative to positive valence), and that the overall impression exhibited by the virtual audience can be controlled my manipulating the amount of individual audience members that display a congruent state.
772.
Reger, Greg M.; Koenen-Woods, Patricia; Zetocha, Kimberlee; Smolenski, Derek J.; Holloway, Kevin M.; Rothbaum, Barbara O.; Difede, JoAnn; Rizzo, Albert A.; Edwards-Stewart, Amanda; Skopp, Nancy A.; Mishkind, Matthew; Reger, Mark A.; Gahm, Gregory A.
In: Journal of Consulting and Clinical Psychology, 2016, ISSN: 1939-2117, 0022-006X.
@article{reger_randomized_2016,
title = {Randomized Controlled Trial of Prolonged Exposure Using Imaginal Exposure vs. Virtual Reality Exposure in Active Duty Soldiers With Deployment-Related Posttraumatic Stress Disorder (PTSD).},
author = {Greg M. Reger and Patricia Koenen-Woods and Kimberlee Zetocha and Derek J. Smolenski and Kevin M. Holloway and Barbara O. Rothbaum and JoAnn Difede and Albert A. Rizzo and Amanda Edwards-Stewart and Nancy A. Skopp and Matthew Mishkind and Mark A. Reger and Gregory A. Gahm},
url = {https://www.researchgate.net/profile/Amanda_Edwards-Stewart/publication/307950241_Randomized_Controlled_Trial_of_Prolonged_Exposure_Using_Imaginal_Exposure_vs_Virtual_Reality_Exposure_in_Active_Duty_Soldiers_With_Deployment-Related_Posttraumatic_Stress_Disorder_PTSD/links/57d6f13f08ae601b39ac25d9.pdf},
doi = {10.1037/ccp0000134},
issn = {1939-2117, 0022-006X},
year = {2016},
date = {2016-09-01},
journal = {Journal of Consulting and Clinical Psychology},
abstract = {Prolonged exposure (PE) is an evidence-based psychotherapy for posttraumatic stress disorder (PTSD) but there is limited research with active-duty military populations. Virtual reality exposure (VRE) has shown promise but randomized trials are needed to evaluate efficacy relative to existing standards of care. This study evaluated the efficacy of VRE and PE for active duty soldiers with PTSD from deployments to Iraq and Afghanistan. Active-duty soldiers ( = 162) were randomized to 10-sessions of PE, VRE, or a minimal attention waitlist (WL). Blinded assessors evaluated symptoms at baseline, halfway through treatment, at posttreatment, and at 3- and 6-month follow-ups using the Clinician Administered PTSD Scale (CAPS). Intent-to-treat analyses found that both PE and VRE resulted in significant reductions in PTSD symptoms relative to those in the WL. The majority of patients demonstrated reliable change in PTSD symptoms. There was no difference between PE and VRE regarding treatment drop out before completing 10 sessions (44 and 41% for VRE and PE, respectively). Contrary to hypotheses, analyses at posttreatment did not show that VRE was superior to PE. Post hoc analyses found that PE resulted in significantly greater symptom reductions than VRE at 3- and 6-month follow-up. Both treatments significantly reduced self-reported stigma. PE is an efficacious treatment for active-duty Army soldiers with PTSD from deployments to Iraq or Afghanistan. Results extend previous evidence supporting the efficacy of PE to active-duty military personnel and raise important questions for future research on VRE},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Prolonged exposure (PE) is an evidence-based psychotherapy for posttraumatic stress disorder (PTSD) but there is limited research with active-duty military populations. Virtual reality exposure (VRE) has shown promise but randomized trials are needed to evaluate efficacy relative to existing standards of care. This study evaluated the efficacy of VRE and PE for active duty soldiers with PTSD from deployments to Iraq and Afghanistan. Active-duty soldiers ( = 162) were randomized to 10-sessions of PE, VRE, or a minimal attention waitlist (WL). Blinded assessors evaluated symptoms at baseline, halfway through treatment, at posttreatment, and at 3- and 6-month follow-ups using the Clinician Administered PTSD Scale (CAPS). Intent-to-treat analyses found that both PE and VRE resulted in significant reductions in PTSD symptoms relative to those in the WL. The majority of patients demonstrated reliable change in PTSD symptoms. There was no difference between PE and VRE regarding treatment drop out before completing 10 sessions (44 and 41% for VRE and PE, respectively). Contrary to hypotheses, analyses at posttreatment did not show that VRE was superior to PE. Post hoc analyses found that PE resulted in significantly greater symptom reductions than VRE at 3- and 6-month follow-up. Both treatments significantly reduced self-reported stigma. PE is an efficacious treatment for active-duty Army soldiers with PTSD from deployments to Iraq or Afghanistan. Results extend previous evidence supporting the efficacy of PE to active-duty military personnel and raise important questions for future research on VRE
773.
Hiraoka, Takuya; Georgila, Kallirroi; Nouri, Elnaz; Traum, David; Nakamura, Satoshi
Reinforcement Learning of Multi-Party Trading Dialog Policies Journal Article
In: Transactions of the Japanese Society for Artificial Intelligence, vol. 31, 2016, ISSN: 1346-8030.
@article{hiraoka_reinforcement_2016,
title = {Reinforcement Learning of Multi-Party Trading Dialog Policies},
author = {Takuya Hiraoka and Kallirroi Georgila and Elnaz Nouri and David Traum and Satoshi Nakamura},
url = {https://www.jstage.jst.go.jp/article/tjsai/31/4/31_B-FC1/_pdf},
issn = {1346-8030},
year = {2016},
date = {2016-09-01},
journal = {Transactions of the Japanese Society for Artificial Intelligence},
volume = {31},
abstract = {Trading dialogs are a kind of negotiation in which an exchange of ownership of items is discussed, and these kinds of dialogs are pervasive in many situations. Recently, there has been an increasing amount of research on applying reinforcement learning (RL) to negotiation dialog domains. However, in previous research, the focus was on negotiation dialog between two participants only, ignoring cases where negotiation takes place between more than two interlocutors. In this paper, as a first study on multi-party negotiation, we apply RL to a multi-party trading scenario where the dialog system (learner) trades with one, two, or three other agents. We experiment with different RL algorithms and reward functions. We use Q-learning with linear function approximation, least-squares policy iteration, and neural fitted Q iteration. In addition, to make the learning process more efficient, we introduce an incremental reward function. The negotiation strategy of the learner is learned through simulated dialog with trader simulators. In our experiments, we evaluate how the performance of the learner varies depending on the RL algorithm used and the number of traders. Furthermore, we compare the learned dialog policies with two strong hand-crafted baseline dialog policies. Our results show that (1) even in simple multi-party trading dialog tasks, learning an effective negotiation policy is not a straightforward task and requires a lot of experimentation; and (2) the use of neural fitted Q iteration combined with an incremental reward function produces negotiation policies as effective or even better than the policies of the two strong hand-crafted baselines.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Trading dialogs are a kind of negotiation in which an exchange of ownership of items is discussed, and these kinds of dialogs are pervasive in many situations. Recently, there has been an increasing amount of research on applying reinforcement learning (RL) to negotiation dialog domains. However, in previous research, the focus was on negotiation dialog between two participants only, ignoring cases where negotiation takes place between more than two interlocutors. In this paper, as a first study on multi-party negotiation, we apply RL to a multi-party trading scenario where the dialog system (learner) trades with one, two, or three other agents. We experiment with different RL algorithms and reward functions. We use Q-learning with linear function approximation, least-squares policy iteration, and neural fitted Q iteration. In addition, to make the learning process more efficient, we introduce an incremental reward function. The negotiation strategy of the learner is learned through simulated dialog with trader simulators. In our experiments, we evaluate how the performance of the learner varies depending on the RL algorithm used and the number of traders. Furthermore, we compare the learned dialog policies with two strong hand-crafted baseline dialog policies. Our results show that (1) even in simple multi-party trading dialog tasks, learning an effective negotiation policy is not a straightforward task and requires a lot of experimentation; and (2) the use of neural fitted Q iteration combined with an incremental reward function produces negotiation policies as effective or even better than the policies of the two strong hand-crafted baselines.
774.
Bresnahan, T.; Rizzo, A.; Burke, S. L.; Partin, M.; Ahlness, R. M.; Trimmer, M.
Using Virtual Interactive Training Agents (VITA) with Adults with Autism and other Developmental Disabilities Proceedings Article
In: Proceedings of the 2016 International Conference on Disability, Virtual Reality, and Associated Technology, pp. 49–56, ICDVRAT and the University of Reading, Los Angeles, CA, 2016, ISBN: 978-0-7049-1547-3.
@inproceedings{bresnahan_using_2016,
title = {Using Virtual Interactive Training Agents (VITA) with Adults with Autism and other Developmental Disabilities},
author = {T. Bresnahan and A. Rizzo and S. L. Burke and M. Partin and R. M. Ahlness and M. Trimmer},
url = {http://www.icdvrat.org/2016/papers/ICDVRAT2016_S02N2_Bresnahan_etal.pdf},
isbn = {978-0-7049-1547-3},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the 2016 International Conference on Disability, Virtual Reality, and Associated Technology},
pages = {49–56},
publisher = {ICDVRAT and the University of Reading},
address = {Los Angeles, CA},
abstract = {Conversational Virtual Human (VH) agents are increasingly being used to support role-play experiential learning across a range of use-cases and populations. This project examined whether use of the Virtual Interactive Training Agent (VITA) system would improve job interviewing skills in a sample of persons with autism or other developmental disability. The study examined performance differences between baseline and final interviews in face-to-face and virtual reality conditions, and whether statistically significant increases were demonstrated between interviewing conditions. Paired samples t-tests were utilized to examine mean changes in performance by interview stage and in the overall difference between baseline and final interview stages. The preliminary results indicated that VITA is a positive factor when preparing young adults with autism or other developmental disability for employment interviews. Statistically significant results were demonstrated across all pilot conditions and in all but one post-assessment condition.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Conversational Virtual Human (VH) agents are increasingly being used to support role-play experiential learning across a range of use-cases and populations. This project examined whether use of the Virtual Interactive Training Agent (VITA) system would improve job interviewing skills in a sample of persons with autism or other developmental disability. The study examined performance differences between baseline and final interviews in face-to-face and virtual reality conditions, and whether statistically significant increases were demonstrated between interviewing conditions. Paired samples t-tests were utilized to examine mean changes in performance by interview stage and in the overall difference between baseline and final interview stages. The preliminary results indicated that VITA is a positive factor when preparing young adults with autism or other developmental disability for employment interviews. Statistically significant results were demonstrated across all pilot conditions and in all but one post-assessment condition.
775.
Mizukami, Masahiro; Yoshino, Koichiro; Neubig, Graham; Traum, David; Nakamura, Satoshi
Analyzing the Effect of Entrainment on Dialogue Acts Proceedings Article
In: Proceedings of the SIGDIAL 2016 Conference, pp. 310–318, Association for Computational Linguistics, Los Angeles, CA, 2016.
@inproceedings{mizukami_analyzing_2016,
title = {Analyzing the Effect of Entrainment on Dialogue Acts},
author = {Masahiro Mizukami and Koichiro Yoshino and Graham Neubig and David Traum and Satoshi Nakamura},
url = {http://www.sigdial.org/workshops/conference17/proceedings/pdf/SIGDIAL40.pdf},
year = {2016},
date = {2016-09-01},
booktitle = {Proceedings of the SIGDIAL 2016 Conference},
pages = {310–318},
publisher = {Association for Computational Linguistics},
address = {Los Angeles, CA},
abstract = {Entrainment is a factor in dialogue that affects not only human-human but also human-machine interaction. While entrainment on the lexical level is well documented, less is known about how entrainment affects dialogue on a more abstract, structural level. In this paper, we investigate the effect of entrainment on dialogue acts and on lexical choice given dialogue acts, as well as how entrainment changes during a dialogue. We also define a novel measure of entrainment to measure these various types of entrainment. These results may serve as guidelines for dialogue systems that would like to entrain with users in a similar manner.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Entrainment is a factor in dialogue that affects not only human-human but also human-machine interaction. While entrainment on the lexical level is well documented, less is known about how entrainment affects dialogue on a more abstract, structural level. In this paper, we investigate the effect of entrainment on dialogue acts and on lexical choice given dialogue acts, as well as how entrainment changes during a dialogue. We also define a novel measure of entrainment to measure these various types of entrainment. These results may serve as guidelines for dialogue systems that would like to entrain with users in a similar manner.
776.
Krämer, Nicole C.; Karacora, Bilge; Lucas, Gale; Dehghani, Morteza; Rüther, Gina; Gratch, Jonathan
In: Computers & Education, vol. 99, pp. 1–13, 2016, ISSN: 03601315.
@article{kramer_closing_2016,
title = {Closing the gender gap in STEM with friendly male instructors? On the effects of rapport behavior and gender of a virtual agent in an instructional interaction},
author = {Nicole C. Krämer and Bilge Karacora and Gale Lucas and Morteza Dehghani and Gina Rüther and Jonathan Gratch},
url = {http://linkinghub.elsevier.com/retrieve/pii/S0360131516300835},
doi = {10.1016/j.compedu.2016.04.002},
issn = {03601315},
year = {2016},
date = {2016-08-01},
journal = {Computers & Education},
volume = {99},
pages = {1–13},
abstract = {While numerous research endeavors address the effects of pedagogical agents, the role of the agent´s gender and its rapport behavior has been neglected. We hypothesize that a minimal amount of behavioral realism induced by display of rapport is necessary for any social effects to occur in human-computer interaction. Further, in line with results from STEM research on female role models, we assume that especially for female learners a same sex agent will be beneficial. In a 2(student gender)x2(agent gender)x2(rapport behavior yes/no) between subjects design, we investigate whether virtual agents can help enhance participants’ performance, effort and motivation in mathematics. Female and male participants (N = 128) interacted with a male or female virtual agent that either displayed rapport or no rapport. Our results confirm the expected main effect of rapport. However, against expectations, our results do not support the assumption that a same sex agent is beneficial for female learners. Participants’ performance and effort were significantly enhanced when interacting with an agent of opposite gender that displayed rapport. Our results have implications on designing agents for education and training purposes.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
While numerous research endeavors address the effects of pedagogical agents, the role of the agent´s gender and its rapport behavior has been neglected. We hypothesize that a minimal amount of behavioral realism induced by display of rapport is necessary for any social effects to occur in human-computer interaction. Further, in line with results from STEM research on female role models, we assume that especially for female learners a same sex agent will be beneficial. In a 2(student gender)x2(agent gender)x2(rapport behavior yes/no) between subjects design, we investigate whether virtual agents can help enhance participants’ performance, effort and motivation in mathematics. Female and male participants (N = 128) interacted with a male or female virtual agent that either displayed rapport or no rapport. Our results confirm the expected main effect of rapport. However, against expectations, our results do not support the assumption that a same sex agent is beneficial for female learners. Participants’ performance and effort were significantly enhanced when interacting with an agent of opposite gender that displayed rapport. Our results have implications on designing agents for education and training purposes.
777.
Freed, Michael C.; Novak, Laura A.; Killgore, William D. S.; Rauch, Sheila A. M.; Koehlmoos, Tracey P.; Ginsberg, J. P.; Krupnick, Janice L.; Rizzo, Albert "Skip"; Andrews, Anne; Engel, Charles C.
IRB and Research Regulatory Delays Within the Military Health System: Do They Really Matter? And If So, Why and for Whom? Journal Article
In: The American Journal of Bioethics, vol. 16, no. 8, pp. 30–37, 2016, ISSN: 1526-5161, 1536-0075.
@article{freed_irb_2016,
title = {IRB and Research Regulatory Delays Within the Military Health System: Do They Really Matter? And If So, Why and for Whom?},
author = {Michael C. Freed and Laura A. Novak and William D. S. Killgore and Sheila A. M. Rauch and Tracey P. Koehlmoos and J. P. Ginsberg and Janice L. Krupnick and Albert "Skip" Rizzo and Anne Andrews and Charles C. Engel},
url = {http://www.tandfonline.com/doi/full/10.1080/15265161.2016.1187212},
doi = {10.1080/15265161.2016.1187212},
issn = {1526-5161, 1536-0075},
year = {2016},
date = {2016-08-01},
journal = {The American Journal of Bioethics},
volume = {16},
number = {8},
pages = {30–37},
abstract = {Institutional review board (IRB) delays may hinder the successful completion of federally funded research in the U.S. military. When this happens, time-sensitive, mission-relevant questions go unanswered. Research participants face unnecessary burdens and risks if delays squeeze recruitment timelines, resulting in inadequate sample sizes for definitive analyses. More broadly, military members are exposed to untested or undertested interventions, implemented by well-intentioned leaders who bypass the research process altogether. To illustrate, we offer two case examples. We posit that IRB delays often appear in the service of managing institutional risk, rather than protecting research participants. Regulators may see more risk associated with moving quickly than risk related to delay, choosing to err on the side of bureaucracy. The authors of this article, all of whom are military-funded researchers, government stakeholders, and/or human subject protection experts, offer feasible recommendations to improve the IRB system and, ultimately, research within military, veteran, and civilian populations.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Institutional review board (IRB) delays may hinder the successful completion of federally funded research in the U.S. military. When this happens, time-sensitive, mission-relevant questions go unanswered. Research participants face unnecessary burdens and risks if delays squeeze recruitment timelines, resulting in inadequate sample sizes for definitive analyses. More broadly, military members are exposed to untested or undertested interventions, implemented by well-intentioned leaders who bypass the research process altogether. To illustrate, we offer two case examples. We posit that IRB delays often appear in the service of managing institutional risk, rather than protecting research participants. Regulators may see more risk associated with moving quickly than risk related to delay, choosing to err on the side of bureaucracy. The authors of this article, all of whom are military-funded researchers, government stakeholders, and/or human subject protection experts, offer feasible recommendations to improve the IRB system and, ultimately, research within military, veteran, and civilian populations.
778.
Washburn, Micki; Bordnick, Patrick; Rizzo, Albert “Skip”
A pilot feasibility study of virtual patient simulation to enhance social work students’ brief mental health assessment skills Journal Article
In: Social Work in Health Care, pp. 1–19, 2016, ISSN: 0098-1389, 1541-034X.
@article{washburn_pilot_2016,
title = {A pilot feasibility study of virtual patient simulation to enhance social work students’ brief mental health assessment skills},
author = {Micki Washburn and Patrick Bordnick and Albert “Skip” Rizzo},
url = {https://www.tandfonline.com/doi/full/10.1080/00981389.2016.1210715},
doi = {10.1080/00981389.2016.1210715},
issn = {0098-1389, 1541-034X},
year = {2016},
date = {2016-08-01},
journal = {Social Work in Health Care},
pages = {1–19},
abstract = {This study presents preliminary feasibility and acceptability data on the use of virtual patient (VP) simulations to develop brief assessment skills within an interdisciplinary care setting. Results support the acceptability of technology-enhanced simulations and offer preliminary evidence for an association between engagement in VP practice simulations and improvements in diagnostic accuracy and clinical interviewing skills. Recommendations and next steps for research on technologyenhanced simulations within social work are discussed.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This study presents preliminary feasibility and acceptability data on the use of virtual patient (VP) simulations to develop brief assessment skills within an interdisciplinary care setting. Results support the acceptability of technology-enhanced simulations and offer preliminary evidence for an association between engagement in VP practice simulations and improvements in diagnostic accuracy and clinical interviewing skills. Recommendations and next steps for research on technologyenhanced simulations within social work are discussed.
779.
Nye, Benjamin D.; Hu, Xiangen
Conceptualizing and Representing Domains to Guide Tutoring Book Section
In: Design Recommendations for Intelligent Tutoring Systems: Volume 4-Domain Modeling, vol. 4, pp. 15–18, US Army Research Laboratory, Orlando, FL, 2016.
@incollection{nye_conceptualizing_2016,
title = {Conceptualizing and Representing Domains to Guide Tutoring},
author = {Benjamin D. Nye and Xiangen Hu},
url = {http://books.google.com/books?hl=en&lr=&id=0suvDAAAQBAJ&oi=fnd&pg=PA15&dq=%22data.+This+chapter+presents+an+excellent+overview+of+current+research+on+Q-matrices%22+%22edge+work+on+ensemble+methods+that+achieve+state+of+the+art+performance+by+combining%22+&ots=6MJhm1XHVV&sig=i14eJyin69Cy-jms2lWIFF4K3CU},
year = {2016},
date = {2016-07-01},
booktitle = {Design Recommendations for Intelligent Tutoring Systems: Volume 4-Domain Modeling},
volume = {4},
pages = {15–18},
publisher = {US Army Research Laboratory},
address = {Orlando, FL},
abstract = {Any discussion about how intelligent tutoring system (ITS) domains must begin with considering how ITS conceptualize and represent domains. This process requires building formal, mathematically-specifiable operationalization of the often implicit knowledge about learning domains and their pedagogy. Across different domains and pedagogical approaches, a wide variety of methods have been taken: a scope that would be better-covered by an encyclopedia rather than a single book. Since this section could not possibly cover every possible approach to domain modeling, the chapters within this section were instead chosen to cover a representative range of fundamentally-different approaches to domain modeling.},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Any discussion about how intelligent tutoring system (ITS) domains must begin with considering how ITS conceptualize and represent domains. This process requires building formal, mathematically-specifiable operationalization of the often implicit knowledge about learning domains and their pedagogy. Across different domains and pedagogical approaches, a wide variety of methods have been taken: a scope that would be better-covered by an encyclopedia rather than a single book. Since this section could not possibly cover every possible approach to domain modeling, the chapters within this section were instead chosen to cover a representative range of fundamentally-different approaches to domain modeling.
780.
Nye, Benjamin D.; Boyce, Michael W.; Sottilare, Robert
Defining the Ill-Defined: From Abstract Principles to Applied Pedagogy Book Section
In: Design Recommendations for Intelligent Tutoring Systems: Volume 4-Domain Modeling, vol. 4, pp. 19–37, US Army Research Laboratory, Orlando, FL, 2016, ISBN: 978-0-9893923-9-6.
@incollection{nye_defining_2016,
title = {Defining the Ill-Defined: From Abstract Principles to Applied Pedagogy},
author = {Benjamin D. Nye and Michael W. Boyce and Robert Sottilare},
url = {https://gifttutoring.org/attachments/download/1736/Design%20Recommendations%20for%20ITS_Volume%204%20-%20Domain%20Modeling%20Book_web%20version_final.pdf},
isbn = {978-0-9893923-9-6},
year = {2016},
date = {2016-07-01},
booktitle = {Design Recommendations for Intelligent Tutoring Systems: Volume 4-Domain Modeling},
volume = {4},
pages = {19–37},
publisher = {US Army Research Laboratory},
address = {Orlando, FL},
abstract = {Attempts to define ill-defined domains in intelligent tutoring system (ITS) research has been approached a number of times (Fournier-Viger, Nkambou, & Nguifo, 2010; Lynch, Ashley, Pinkwart, & Aleven, 2009; Mitrovic & Weerasinghe, 2009; Jacovina, Snow, Dai, & McNamara, 2015; Woods, Stensrud, Wray, Haley, & Jones, 2015). Related research has tried to determine levels of ill-definedness for a domain (Le, Loll, & Pinkwart, 2013). Despite such attempts, the field has not yet converged on common guidelines to distinguish between well-defined versus ill-defined domains. We argue that such guidelines struggle to converge because a domain is too large to meaningfully categorize: every domain contains a mixture of well-defined and ill-defined tasks. While the co-existence of well-defined and ill-defined tasks in a single domain is nearly universally-agreed upon by researchers; this key point is often quickly buried by an extensive discussion about what makes certain domain tasks ill-defined (e.g., disagreement about ideal solutions, multiple solution paths). In this chapter, we first take a step back to consider what is meant by a domain in the context of learning. Next, based on this definition for a domain, we map out the components that are in a learning domain, since each component may have ill-defined parts. This leads into a discussion about the strategies that have been used to make ill-defined domains tractable for certain types of pedagogy. Examples of ITS research that applies these strategies are noted. Finally, we conclude with practical how-to considerations and open research questions for approaching ill-defined domains. This chapter should be considered a companion piece to our chapter in the prior volume of this series (Nye, Goldberg, & Hu, 2015). This chapter focuses on how to understand and transform ill-defined parts of domains, while the prior chapter discusses commonly-used learning tasks and authoring approaches for both well-defined and ill-defined tasks. As such, this chapter is intended to help the learner understand if and how different parts of the domain are ill-defined (and what to do about them). The companion piece in the authoring tools volume discusses different categories of well and ill-defined tasks, from the standpoint of attempting to author and maintain an ITS.},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Attempts to define ill-defined domains in intelligent tutoring system (ITS) research has been approached a number of times (Fournier-Viger, Nkambou, & Nguifo, 2010; Lynch, Ashley, Pinkwart, & Aleven, 2009; Mitrovic & Weerasinghe, 2009; Jacovina, Snow, Dai, & McNamara, 2015; Woods, Stensrud, Wray, Haley, & Jones, 2015). Related research has tried to determine levels of ill-definedness for a domain (Le, Loll, & Pinkwart, 2013). Despite such attempts, the field has not yet converged on common guidelines to distinguish between well-defined versus ill-defined domains. We argue that such guidelines struggle to converge because a domain is too large to meaningfully categorize: every domain contains a mixture of well-defined and ill-defined tasks. While the co-existence of well-defined and ill-defined tasks in a single domain is nearly universally-agreed upon by researchers; this key point is often quickly buried by an extensive discussion about what makes certain domain tasks ill-defined (e.g., disagreement about ideal solutions, multiple solution paths). In this chapter, we first take a step back to consider what is meant by a domain in the context of learning. Next, based on this definition for a domain, we map out the components that are in a learning domain, since each component may have ill-defined parts. This leads into a discussion about the strategies that have been used to make ill-defined domains tractable for certain types of pedagogy. Examples of ITS research that applies these strategies are noted. Finally, we conclude with practical how-to considerations and open research questions for approaching ill-defined domains. This chapter should be considered a companion piece to our chapter in the prior volume of this series (Nye, Goldberg, & Hu, 2015). This chapter focuses on how to understand and transform ill-defined parts of domains, while the prior chapter discusses commonly-used learning tasks and authoring approaches for both well-defined and ill-defined tasks. As such, this chapter is intended to help the learner understand if and how different parts of the domain are ill-defined (and what to do about them). The companion piece in the authoring tools volume discusses different categories of well and ill-defined tasks, from the standpoint of attempting to author and maintain an ITS.
Filter
2006
Woods, Steven Paul; Rippeth, Julie D.; Conover, Emily; Carey, Catherine L.; Parsons, Thomas D.; Tröster, Alexander I.
Statistical Power of Studies Examining the Cognitive Effects of Subthalamic Nucleus Deep Brain Stimulation in Parkinson's Disease Journal Article
In: The Clinical Neuropsychologist, vol. 20, pp. 27–38, 2006, ISSN: 1385-4046.
Abstract | Links | BibTeX | Tags: MedVR
@article{woods_statistical_2006,
title = {Statistical Power of Studies Examining the Cognitive Effects of Subthalamic Nucleus Deep Brain Stimulation in Parkinson's Disease},
author = {Steven Paul Woods and Julie D. Rippeth and Emily Conover and Catherine L. Carey and Thomas D. Parsons and Alexander I. Tröster},
url = {http://ict.usc.edu/pubs/STATISTICAL%20POWER%20OF%20STUDIES%20EXAMINING%20THE%20COGNITIVE%20EFFECTS%20OF%20SUBTHALAMIC%20NUCLEUS%20DEEP%20BRAIN%20STIMULATION%20IN%20PARKINSON%E2%80%99S%20DISEASE.pdf},
doi = {10.1080/13854040500203290},
issn = {1385-4046},
year = {2006},
date = {2006-01-01},
journal = {The Clinical Neuropsychologist},
volume = {20},
pages = {27–38},
abstract = {It has been argued that neuropsychological studies generally possess adequate statistical power to detect large effect sizes. However, low statistical power is problematic in neuropsychological research involving clinical populations and novel interventions for which available sample sizes are often limited. One notable example of this problem is evident in the literature regarding the cognitive sequelae of deep brain stimulation (DBS) of the subthalamic nucleus (STN) in persons with Parkinson's disease (PD). In the current review, a post hoc estimate of the statistical power of 30 studies examining cognitive effects of STN DBS in PD revealed adequate power to detect substantial cognitive declines (i.e., very large effect sizes), but surprisingly low estimated power to detect cognitive changes associated with conventionally small, medium, and large effect sizes. Such wide spread Type II error risk in the STN DBS cognitive outcomes literature may affect the clinical decision-making process as concerns the possible risk of postsurgical cognitive morbidity, as well as conceptual inferences to be drawn regarding the role of the STN in higher-level cognitive functions. Statistical and methodological recommendations (e.g., meta-analysis) are offered to enhance the power of current and future studies examining the neuropsychological sequelae of STN DBS in PD.},
keywords = {MedVR},
pubstate = {published},
tppubtype = {article}
}
It has been argued that neuropsychological studies generally possess adequate statistical power to detect large effect sizes. However, low statistical power is problematic in neuropsychological research involving clinical populations and novel interventions for which available sample sizes are often limited. One notable example of this problem is evident in the literature regarding the cognitive sequelae of deep brain stimulation (DBS) of the subthalamic nucleus (STN) in persons with Parkinson's disease (PD). In the current review, a post hoc estimate of the statistical power of 30 studies examining cognitive effects of STN DBS in PD revealed adequate power to detect substantial cognitive declines (i.e., very large effect sizes), but surprisingly low estimated power to detect cognitive changes associated with conventionally small, medium, and large effect sizes. Such wide spread Type II error risk in the STN DBS cognitive outcomes literature may affect the clinical decision-making process as concerns the possible risk of postsurgical cognitive morbidity, as well as conceptual inferences to be drawn regarding the role of the STN in higher-level cognitive functions. Statistical and methodological recommendations (e.g., meta-analysis) are offered to enhance the power of current and future studies examining the neuropsychological sequelae of STN DBS in PD.
Martinovski, Bilyana
Framework for analysis of mitigation in courts Journal Article
In: Journal of Pragmatics, 2006.
Abstract | Links | BibTeX | Tags:
@article{martinovski_framework_2006,
title = {Framework for analysis of mitigation in courts},
author = {Bilyana Martinovski},
url = {http://ict.usc.edu/pubs/Framework%20for%20analysis%20of%20mitigation%20in%20courts.pdf},
year = {2006},
date = {2006-01-01},
journal = {Journal of Pragmatics},
abstract = {This paper presents an activity-based framework for empirical discourse analysis of mitigation in public environments such as Swedish and Bulgarian courtroom examinations. Mitigation is defined as a pragmatic, cognitive and linguistic behavior the main purpose of which is reduction of vulnerability. The suggested framework consists of mitigation processes, which involve mitigating argumentation lines, defense moves, and communicative acts. The functions of mitigation are described in terms of the participants' actions and goals separately from politeness strategies. The conclusions and observations address two things: issues related to the pragmatic theory of communication especially mitigation and issues related to the trial as a social activity. For instance, non-turn-taking confirmations by examiners are often followed by volunteered utterances, which in some cases may be examples of 'rehearsed' testimonies. At the same time the witnesses' tendency to volunteer information even on the behalf of their own credibility indicates that they also favor pro-party testimonies. Despite the objective judicial role of the prosecutor or judge and/or despite the examiners accommodating style the verbal behavior of the witnesses exhibits constant anticipation of danger.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
This paper presents an activity-based framework for empirical discourse analysis of mitigation in public environments such as Swedish and Bulgarian courtroom examinations. Mitigation is defined as a pragmatic, cognitive and linguistic behavior the main purpose of which is reduction of vulnerability. The suggested framework consists of mitigation processes, which involve mitigating argumentation lines, defense moves, and communicative acts. The functions of mitigation are described in terms of the participants' actions and goals separately from politeness strategies. The conclusions and observations address two things: issues related to the pragmatic theory of communication especially mitigation and issues related to the trial as a social activity. For instance, non-turn-taking confirmations by examiners are often followed by volunteered utterances, which in some cases may be examples of 'rehearsed' testimonies. At the same time the witnesses' tendency to volunteer information even on the behalf of their own credibility indicates that they also favor pro-party testimonies. Despite the objective judicial role of the prosecutor or judge and/or despite the examiners accommodating style the verbal behavior of the witnesses exhibits constant anticipation of danger.
Leuski, Anton; Pair, Jarrell; Traum, David; McNerney, Peter J.; Georgiou, Panayiotis G.; Patel, Ronakkumar
How to Talk to a Hologram Proceedings Article
In: Proceedings of the 11th International Conference on Intelligent User Interfaces, Sydney, Australia, 2006.
Abstract | Links | BibTeX | Tags: Virtual Humans
@inproceedings{leuski_how_2006,
title = {How to Talk to a Hologram},
author = {Anton Leuski and Jarrell Pair and David Traum and Peter J. McNerney and Panayiotis G. Georgiou and Ronakkumar Patel},
url = {http://ict.usc.edu/pubs/How%20to%20Talk%20to%20a%20Hologram.pdf},
year = {2006},
date = {2006-01-01},
booktitle = {Proceedings of the 11th International Conference on Intelligent User Interfaces},
address = {Sydney, Australia},
abstract = {There is a growing need for creating life-like virtual human simulations that can conduct a natural spoken dialog with a human student on a predefined subject. We present an overview of a spoken-dialog system that supports a person interacting with a full-size hologram-like virtual human character in an exhibition kiosk settings. We also give a brief summary of the natural language classification component of the system and describe the experiments we conducted with the system.},
keywords = {Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
There is a growing need for creating life-like virtual human simulations that can conduct a natural spoken dialog with a human student on a predefined subject. We present an overview of a spoken-dialog system that supports a person interacting with a full-size hologram-like virtual human character in an exhibition kiosk settings. We also give a brief summary of the natural language classification component of the system and describe the experiments we conducted with the system.
Parsons, Thomas D.; Braaten, Alyssa J.; Hall, Colin D.; Robertson, R. Kevin
Better quality of life with neuropsychological improvement on HAART Journal Article
In: Health and Quality of Life Outcomes, vol. 4, no. 11, 2006.
Abstract | Links | BibTeX | Tags: MedVR
@article{parsons_better_2006,
title = {Better quality of life with neuropsychological improvement on HAART},
author = {Thomas D. Parsons and Alyssa J. Braaten and Colin D. Hall and R. Kevin Robertson},
url = {http://ict.usc.edu/pubs/Better%20quality%20of%20life%20with%20neuropsychological%20improvement%20on%20HAART.pdf},
year = {2006},
date = {2006-01-01},
journal = {Health and Quality of Life Outcomes},
volume = {4},
number = {11},
abstract = {Background: Successful highly active antiretroviral therapy (HAART) regimens have resulted in substantial improvements in the systemic health of HIV infected persons and increased survival times. Despite increased systemic health, the prevalence of minor HIV-associated cognitive impairment appears to be rising with increased longevity, and it remains to be seen what functional outcomes will result from these improvements. Cognitive impairment can dramatically impact functional ability and day-to-day productivity. We assessed the relationship of quality of life (QOL) and neuropsychological functioning with successful HAART treatment. Methods: In a prospective longitudinal study, subjects were evaluated before instituting HAART (naïve) or before changing HAART regimens because current therapy failed to maintain suppression of plasma viral load (treatment failure). Subjects underwent detailed neuropsychological and neurological examinations, as well as psychological evaluation sensitive to possible confounds. Re-evaluation was performed six months after institution of the new HAART regimen and/or if plasma viral load indicated treatment failure. At each evaluation, subjects underwent ultrasensitive HIV RNA quantitative evaluation in both plasma and cerebrospinal fluid. Results: HAART successes performed better than failures on measures exploring speed of mental processing (p textbackslashtextbackslashtextbackslashtextbackslashtextless .02). HAART failure was significantly associated with increased self-reports of physical health complaints (p textbackslashtextbackslashtextbackslashtextbackslashtextless .01) and substance abuse (p textbackslashtextbackslashtextbackslashtextbackslashtextless .01). An interesting trend emerged, in which HAART failures endorsed greater levels of psychological and cognitive complaints (p = 06). Analysis between neuropsychological measures and QOL scores revealed significant. correlation between QOL Total and processing speed (p textbackslashtextbackslashtextbackslashtextbackslashtextless .05), as well as flexibility (p textbackslashtextbackslashtextbackslashtextbackslashtextless .05). Conclusion: Our study investigated the relationship between HIV-associated neurocognitive impairment and quality of life. HAART failures experienced slower psychomotor processing, and had increased self-reports of physical health complaints and substance abuse. Contrariwise, HAART successes experienced improved mental processing, demonstrating the impact of successful treatment on functioning. With increasing life expectancy for those who are HIV seropositive, it is important to measure cognitive functioning in relation to the actual QOL these individuals report. The study results have implications for the optimal management of HIV-infected persons. Specific support or intervention may be beneficial for those who have failed HAART in order to decrease substance abuse and increase overall physical health.},
keywords = {MedVR},
pubstate = {published},
tppubtype = {article}
}
Background: Successful highly active antiretroviral therapy (HAART) regimens have resulted in substantial improvements in the systemic health of HIV infected persons and increased survival times. Despite increased systemic health, the prevalence of minor HIV-associated cognitive impairment appears to be rising with increased longevity, and it remains to be seen what functional outcomes will result from these improvements. Cognitive impairment can dramatically impact functional ability and day-to-day productivity. We assessed the relationship of quality of life (QOL) and neuropsychological functioning with successful HAART treatment. Methods: In a prospective longitudinal study, subjects were evaluated before instituting HAART (naïve) or before changing HAART regimens because current therapy failed to maintain suppression of plasma viral load (treatment failure). Subjects underwent detailed neuropsychological and neurological examinations, as well as psychological evaluation sensitive to possible confounds. Re-evaluation was performed six months after institution of the new HAART regimen and/or if plasma viral load indicated treatment failure. At each evaluation, subjects underwent ultrasensitive HIV RNA quantitative evaluation in both plasma and cerebrospinal fluid. Results: HAART successes performed better than failures on measures exploring speed of mental processing (p textbackslashtextbackslashtextbackslashtextbackslashtextless .02). HAART failure was significantly associated with increased self-reports of physical health complaints (p textbackslashtextbackslashtextbackslashtextbackslashtextless .01) and substance abuse (p textbackslashtextbackslashtextbackslashtextbackslashtextless .01). An interesting trend emerged, in which HAART failures endorsed greater levels of psychological and cognitive complaints (p = 06). Analysis between neuropsychological measures and QOL scores revealed significant. correlation between QOL Total and processing speed (p textbackslashtextbackslashtextbackslashtextbackslashtextless .05), as well as flexibility (p textbackslashtextbackslashtextbackslashtextbackslashtextless .05). Conclusion: Our study investigated the relationship between HIV-associated neurocognitive impairment and quality of life. HAART failures experienced slower psychomotor processing, and had increased self-reports of physical health complaints and substance abuse. Contrariwise, HAART successes experienced improved mental processing, demonstrating the impact of successful treatment on functioning. With increasing life expectancy for those who are HIV seropositive, it is important to measure cognitive functioning in relation to the actual QOL these individuals report. The study results have implications for the optimal management of HIV-infected persons. Specific support or intervention may be beneficial for those who have failed HAART in order to decrease substance abuse and increase overall physical health.
Gratch, Jonathan; Mao, Wenji; Marsella, Stacy C.
Modeling Social Emotions and Social Attributions Book Section
In: Sun, R. (Ed.): Cognition and Multi-Agent Interaction: Extending Cognitive Modeling to Social Simulation, Cambridge University Press, 2006.
Links | BibTeX | Tags: Social Simulation, Virtual Humans
@incollection{gratch_modeling_2006,
title = {Modeling Social Emotions and Social Attributions},
author = {Jonathan Gratch and Wenji Mao and Stacy C. Marsella},
editor = {R. Sun},
url = {http://ict.usc.edu/pubs/Modeling%20Social%20Emotions%20and%20Social%20Attributions.pdf},
year = {2006},
date = {2006-01-01},
booktitle = {Cognition and Multi-Agent Interaction: Extending Cognitive Modeling to Social Simulation},
publisher = {Cambridge University Press},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {incollection}
}
Gold, Jeffrey I.; Kim, Seok Hyeon; Kant, Alexis J.; Joseph, Michael H.; Rizzo, Albert
Effectiveness of Virtual Reality for Pediatric Pain Distraction during IV Placement Journal Article
In: CyberPsychology and Behavior, vol. 9, no. 2, pp. 207–212, 2006.
Abstract | Links | BibTeX | Tags: MedVR
@article{gold_effectiveness_2006,
title = {Effectiveness of Virtual Reality for Pediatric Pain Distraction during IV Placement},
author = {Jeffrey I. Gold and Seok Hyeon Kim and Alexis J. Kant and Michael H. Joseph and Albert Rizzo},
url = {http://ict.usc.edu/pubs/Effectiveness%20of%20Virtual%20Reality%20for%20Pediatric%20Pain%20Distraction%20during%20IV%20Placement.pdf},
year = {2006},
date = {2006-01-01},
journal = {CyberPsychology and Behavior},
volume = {9},
number = {2},
pages = {207–212},
abstract = {The objective of this study was to test the efficacy and suitability of virtual reality (VR) as apain distraction for pediatric intravenous (IV) placement. Twenty children (12 boys, 8 girls) requiring IV placement for a magnetic resonance imaging/computed tomography (MRI/CT) scan were randomly assigned to two conditions: (1) VR distraction using Street Luge(5DT), presented via a head-mounted display, or (2) standard of care (topical anesthetic) with no distraction. Children, their parents, and nurses completed self-report questionnaires that assessed numerous health-related outcomes. Responses from the Faces Pain Scale–Revisedindicated a fourfold increase in affective pain within the control condition; by contrast, nosignificant differences were detected within the VR condition. Significant associations between multiple measures of anticipatory anxiety, affective pain, IV pain intensity, and measures of past procedural pain provided support for the complex interplay of a multimodalassessment of pain perception. There was also a sufficient amount of evidence supportingthe efficacy of Street Luge as a pediatric pain distraction tool during IV placement: an adequate level of presence, no simulator sickness, and significantly more child-, parent-, and nurse-reported satisfaction with pain management. VR pain distraction was positively endorsed by all reporters and is a promising tool for decreasing pain, and anxiety in childrenundergoing acute medical interventions. However, further research with larger sample sizesand other routine medical procedures is warranted.},
keywords = {MedVR},
pubstate = {published},
tppubtype = {article}
}
The objective of this study was to test the efficacy and suitability of virtual reality (VR) as apain distraction for pediatric intravenous (IV) placement. Twenty children (12 boys, 8 girls) requiring IV placement for a magnetic resonance imaging/computed tomography (MRI/CT) scan were randomly assigned to two conditions: (1) VR distraction using Street Luge(5DT), presented via a head-mounted display, or (2) standard of care (topical anesthetic) with no distraction. Children, their parents, and nurses completed self-report questionnaires that assessed numerous health-related outcomes. Responses from the Faces Pain Scale–Revisedindicated a fourfold increase in affective pain within the control condition; by contrast, nosignificant differences were detected within the VR condition. Significant associations between multiple measures of anticipatory anxiety, affective pain, IV pain intensity, and measures of past procedural pain provided support for the complex interplay of a multimodalassessment of pain perception. There was also a sufficient amount of evidence supportingthe efficacy of Street Luge as a pediatric pain distraction tool during IV placement: an adequate level of presence, no simulator sickness, and significantly more child-, parent-, and nurse-reported satisfaction with pain management. VR pain distraction was positively endorsed by all reporters and is a promising tool for decreasing pain, and anxiety in childrenundergoing acute medical interventions. However, further research with larger sample sizesand other routine medical procedures is warranted.
Tortell, Rebecca; Morie, Jacquelyn
Videogame play and the effectiveness of virtual environments for training Proceedings Article
In: Interservice/Industry Training, Simulation and Education Conference (I/ITSEC), 2006.
Abstract | Links | BibTeX | Tags:
@inproceedings{tortell_videogame_2006,
title = {Videogame play and the effectiveness of virtual environments for training},
author = {Rebecca Tortell and Jacquelyn Morie},
url = {http://ict.usc.edu/pubs/Videogame%20play%20and%20the%20effectiveness%20of%20virtual%20environments%20for%20training.pdf},
year = {2006},
date = {2006-01-01},
booktitle = {Interservice/Industry Training, Simulation and Education Conference (I/ITSEC)},
abstract = {The Sensory Environments Evaluation (SEE) project set out to examine the effects of emotional valence of a virtual training scenario on learning and memory. Emotional arousal is well-established as having enhancing effects on memory (McGaugh, 2000). A virtual scenario called DarkCon was created to resemble a night-time reconnaissance mission. Priming of subjects was the first experimental variable. Subjects were randomly assigned to receive their mission briefing in a serious style, suggesting a serious military mission, or in a lighter style, suggesting a fun roleplaying game. The influence of videogame experience was included in analysis of subjects' recall of the environment and of their physiology. In the present study, 34 Army Rangers from Fort Benning, GA underwent the DarkCon mission. Significant effects of priming condition and videogame play were discovered in subjects' recollection of the mission, and in their physiological reactions to highly exciting material. This paper is primarily concerned with the effects of videogame play frequency on subjects' behavior, recall, and physiology. The effects of priming will be cursorily discussed here as they relate to videogame play habits, and explored in more detail on their own in future publications. Directions for future research into the effects of videogame play experience on training are discussed.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The Sensory Environments Evaluation (SEE) project set out to examine the effects of emotional valence of a virtual training scenario on learning and memory. Emotional arousal is well-established as having enhancing effects on memory (McGaugh, 2000). A virtual scenario called DarkCon was created to resemble a night-time reconnaissance mission. Priming of subjects was the first experimental variable. Subjects were randomly assigned to receive their mission briefing in a serious style, suggesting a serious military mission, or in a lighter style, suggesting a fun roleplaying game. The influence of videogame experience was included in analysis of subjects' recall of the environment and of their physiology. In the present study, 34 Army Rangers from Fort Benning, GA underwent the DarkCon mission. Significant effects of priming condition and videogame play were discovered in subjects' recollection of the mission, and in their physiological reactions to highly exciting material. This paper is primarily concerned with the effects of videogame play frequency on subjects' behavior, recall, and physiology. The effects of priming will be cursorily discussed here as they relate to videogame play habits, and explored in more detail on their own in future publications. Directions for future research into the effects of videogame play experience on training are discussed.
2005
Cantzos, Demetrios; Kyriakakis, Chris
Quality Enhancement of Low Bit Rate MPEG1-Layer 3 Audio Based on Audio Resynthesis Proceedings Article
In: Proceedings of the 119th Audio Engineering Society Convention, New York, NY, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{cantzos_quality_2005,
title = {Quality Enhancement of Low Bit Rate MPEG1-Layer 3 Audio Based on Audio Resynthesis},
author = {Demetrios Cantzos and Chris Kyriakakis},
url = {http://ict.usc.edu/pubs/Quality%20Enhancement%20of%20Low%20Bit%20Rate%20MPEG1-Layer%203%20Audio%20Based%20on%20Audio%20Resynthesis.pdf},
year = {2005},
date = {2005-10-01},
booktitle = {Proceedings of the 119th Audio Engineering Society Convention},
address = {New York, NY},
abstract = {One of the most popular audio compression formats is indisputably the MPEG1-Layer 3 format which is based on the idea of low-bit transparent encoding. As these types of audio signals are starting to migrate from portable players with inexpensive headphones to higher quality home audio systems, it is becoming evident that higher bit rates may be required to maintain transparency. We propose a novel method that enhances low bit rate MP3 encoded audio segments by applying multichannel audio resynthesis methods in a post-processing stage or during decoding. Our algorithm employs the highly efficient Generalized Gaussian mixture model which, combined with cepstral smoothing, leads to very low cepstral reconstruction errors. In addition, residual conversion is applied which proves to significantly improve the enhancement performance. The method presented can be easily generalized to include other audio formats for which sound quality is an issue.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
One of the most popular audio compression formats is indisputably the MPEG1-Layer 3 format which is based on the idea of low-bit transparent encoding. As these types of audio signals are starting to migrate from portable players with inexpensive headphones to higher quality home audio systems, it is becoming evident that higher bit rates may be required to maintain transparency. We propose a novel method that enhances low bit rate MP3 encoded audio segments by applying multichannel audio resynthesis methods in a post-processing stage or during decoding. Our algorithm employs the highly efficient Generalized Gaussian mixture model which, combined with cepstral smoothing, leads to very low cepstral reconstruction errors. In addition, residual conversion is applied which proves to significantly improve the enhancement performance. The method presented can be easily generalized to include other audio formats for which sound quality is an issue.
Gordon, Andrew S.; Ganesan, Kavita
Automated Story Capture From Conversational Speech Proceedings Article
In: 3rd International Conference on Knowledge Capture (K-CAP 05), Banff, Alberta, Canada, 2005.
Abstract | Links | BibTeX | Tags: The Narrative Group
@inproceedings{gordon_automated_2005,
title = {Automated Story Capture From Conversational Speech},
author = {Andrew S. Gordon and Kavita Ganesan},
url = {http://ict.usc.edu/pubs/Automated%20Story%20Capture%20From%20Conversational%20Speech.pdf},
year = {2005},
date = {2005-10-01},
booktitle = {3rd International Conference on Knowledge Capture (K-CAP 05)},
address = {Banff, Alberta, Canada},
abstract = {While storytelling has long been recognized as an important part of effective knowledge management in organizations, knowledge management technologies have generally not distinguished between stories and other types of discourse. In this paper we describe a new type of technological support for storytelling that involves automatically capturing the stories that people tell to each other in conversations. We describe our first attempt at constructing an automated story extraction system using statistical text classification and a simple voting scheme. We evaluate the performance of this system and demonstrate that useful levels of precision and recall can be obtained when analyzing transcripts of interviews, but that performance on speech recognition data is not above what can be expected by chance. This paper establishes the level of performance that can be obtained using a straightforward approach to story extraction, and outlines ways in which future systems can improve on these results and enable a wide range of knowledge socialization applications.},
keywords = {The Narrative Group},
pubstate = {published},
tppubtype = {inproceedings}
}
While storytelling has long been recognized as an important part of effective knowledge management in organizations, knowledge management technologies have generally not distinguished between stories and other types of discourse. In this paper we describe a new type of technological support for storytelling that involves automatically capturing the stories that people tell to each other in conversations. We describe our first attempt at constructing an automated story extraction system using statistical text classification and a simple voting scheme. We evaluate the performance of this system and demonstrate that useful levels of precision and recall can be obtained when analyzing transcripts of interviews, but that performance on speech recognition data is not above what can be expected by chance. This paper establishes the level of performance that can be obtained using a straightforward approach to story extraction, and outlines ways in which future systems can improve on these results and enable a wide range of knowledge socialization applications.
Traum, David; Swartout, William; Gratch, Jonathan; Marsella, Stacy C.; Kenny, Patrick G.; Hovy, Eduard; Narayanan, Shrikanth; Fast, Edward; Martinovski, Bilyana; Baghat, Rahul; Robinson, Susan; Marshall, Andrew; Wang, Dagen; Gandhe, Sudeep; Leuski, Anton
Dealing with Doctors: A Virtual Human for Non-team Interaction Proceedings Article
In: 6th SIGdial Conference on Discourse and Dialogue, Lisbon, Portugal, 2005.
Abstract | Links | BibTeX | Tags: Social Simulation, Virtual Humans
@inproceedings{traum_dealing_2005,
title = {Dealing with Doctors: A Virtual Human for Non-team Interaction},
author = {David Traum and William Swartout and Jonathan Gratch and Stacy C. Marsella and Patrick G. Kenny and Eduard Hovy and Shrikanth Narayanan and Edward Fast and Bilyana Martinovski and Rahul Baghat and Susan Robinson and Andrew Marshall and Dagen Wang and Sudeep Gandhe and Anton Leuski},
url = {http://ict.usc.edu/pubs/Dealing%20with%20Doctors.pdf},
year = {2005},
date = {2005-09-01},
booktitle = {6th SIGdial Conference on Discourse and Dialogue},
address = {Lisbon, Portugal},
abstract = {We present a virtual human do tor who an engage in multi-modal negotiation dialogue with people from other organizations. The do tor is part of the SASO-ST system, used for training for non-team intera tions},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a virtual human do tor who an engage in multi-modal negotiation dialogue with people from other organizations. The do tor is part of the SASO-ST system, used for training for non-team intera tions
Traum, David; Swartout, William; Marsella, Stacy C.; Gratch, Jonathan
Fight, Flight, or Negotiate: Believable Strategies for Conversing under Crisis Proceedings Article
In: 5th International Working Conference on Intelligent Virtual Agents, Kos, Greece, 2005.
Abstract | Links | BibTeX | Tags: Social Simulation, Virtual Humans
@inproceedings{traum_fight_2005,
title = {Fight, Flight, or Negotiate: Believable Strategies for Conversing under Crisis},
author = {David Traum and William Swartout and Stacy C. Marsella and Jonathan Gratch},
url = {http://ict.usc.edu/pubs/Fight,%20Flight,%20or%20Negotiate-%20Believable%20Strategies%20for%20Conversing%20under%20Crisis.pdf},
year = {2005},
date = {2005-09-01},
booktitle = {5th International Working Conference on Intelligent Virtual Agents},
address = {Kos, Greece},
abstract = {This paper des ribes a model of onversation strategies implemented in virtual humans designed to help people learn negotiation skills. We motivate and dis uss these strategies and their use to allow a virtual human to engage in omplex adversarial negotiation with a human trainee. Choi e of strategy depends on both the personality of the agent and assessment of the likelihood that the negotiation an be bene ial. Exe ution of strategies an be performed by hoosing spe i dialogue behaviors su h as whether and how to respond to a proposal. Current assessment of the value of the topi , the utility of the strategy, and aÆliation toward the other onversants an be used to dynami ally hange strategies throughout the ourse of a onversation. Examples will be given from the SASO-ST proje t, in whi h a trainee learns to negotiate by intera ting with virtual humans who employ these strategies.},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper des ribes a model of onversation strategies implemented in virtual humans designed to help people learn negotiation skills. We motivate and dis uss these strategies and their use to allow a virtual human to engage in omplex adversarial negotiation with a human trainee. Choi e of strategy depends on both the personality of the agent and assessment of the likelihood that the negotiation an be bene ial. Exe ution of strategies an be performed by hoosing spe i dialogue behaviors su h as whether and how to respond to a proposal. Current assessment of the value of the topi , the utility of the strategy, and aÆliation toward the other onversants an be used to dynami ally hange strategies throughout the ourse of a onversation. Examples will be given from the SASO-ST proje t, in whi h a trainee learns to negotiate by intera ting with virtual humans who employ these strategies.
Morie, Jacquelyn; Iyer, Kumar; Luigi, Donat-Pierre; Williams, Josh; Dozois, Aimee; Rizzo, Albert
Development of a Data Management Tool for Investigating Multivariate Space and Free Will Experiences Journal Article
In: Applied Psychophysiology and Biofeedback, vol. 30, no. 3, pp. 319–331, 2005.
Abstract | Links | BibTeX | Tags: MedVR, Virtual Worlds
@article{morie_development_2005,
title = {Development of a Data Management Tool for Investigating Multivariate Space and Free Will Experiences},
author = {Jacquelyn Morie and Kumar Iyer and Donat-Pierre Luigi and Josh Williams and Aimee Dozois and Albert Rizzo},
url = {http://ict.usc.edu/pubs/Development%20of%20a%20Data%20Management%20Tool%20for%20Investigating%20Multivariate%20Space%20and%20Free%20Will%20Experiences%20in%20Virtual%20Reality.pdf},
year = {2005},
date = {2005-09-01},
journal = {Applied Psychophysiology and Biofeedback},
volume = {30},
number = {3},
pages = {319–331},
abstract = {While achieving realism has been a main goal in making convincing virtual reality (VR) environments, just what constitutes realism is still a question situated firmly in the research domain. VR has become mature enough to be used in therapeutic applications such as clinical exposure therapy with some success. We now need detailed scientific investigations to better understand why VR works for these types of cases, and how it could work for other key applications such as training. Just as in real life, it appears that the factors will be complex and multi-variate, and this plethoric situation presents exceptional challenges to the VR researcher. We would not want to lessen VR’s ability to replicate real world conditions in order to more easily study it, however, for by doing so we may compromise the very qualities that comprise its effectiveness. What is really needed are more robust tools to instrument, organize, and visualize the complex data generated by measurements of participant experiences in a realistic virtual world. We describe here our first study in an ongoing program of effective virtual environment research, the types of data we are dealing with, and a specific tool we have been compelled to create that allows us some measure of control over this data. We call this tool Phloem, after the botanical channels that plants use to transport, support and store nutrients.},
keywords = {MedVR, Virtual Worlds},
pubstate = {published},
tppubtype = {article}
}
While achieving realism has been a main goal in making convincing virtual reality (VR) environments, just what constitutes realism is still a question situated firmly in the research domain. VR has become mature enough to be used in therapeutic applications such as clinical exposure therapy with some success. We now need detailed scientific investigations to better understand why VR works for these types of cases, and how it could work for other key applications such as training. Just as in real life, it appears that the factors will be complex and multi-variate, and this plethoric situation presents exceptional challenges to the VR researcher. We would not want to lessen VR’s ability to replicate real world conditions in order to more easily study it, however, for by doing so we may compromise the very qualities that comprise its effectiveness. What is really needed are more robust tools to instrument, organize, and visualize the complex data generated by measurements of participant experiences in a realistic virtual world. We describe here our first study in an ongoing program of effective virtual environment research, the types of data we are dealing with, and a specific tool we have been compelled to create that allows us some measure of control over this data. We call this tool Phloem, after the botanical channels that plants use to transport, support and store nutrients.
Debevec, Paul
Capturing and Simulating Physically Accurate Illumination in Computer Graphics Proceedings Article
In: 11th Annual Symposium on Frontiers of Engineering, Niskayuna, NY, 2005.
Abstract | Links | BibTeX | Tags: Graphics
@inproceedings{debevec_capturing_2005,
title = {Capturing and Simulating Physically Accurate Illumination in Computer Graphics},
author = {Paul Debevec},
url = {http://ict.usc.edu/pubs/Capturing%20and%20Simulating%20Physically%20Accurate%20Illumination%20in%20Computer%20Graphics.pdf},
year = {2005},
date = {2005-09-01},
booktitle = {11th Annual Symposium on Frontiers of Engineering},
address = {Niskayuna, NY},
abstract = {Anyone who has seen a recent summer blockbuster has witnessed the dramatic increases in computer-generated realism in recent years. Visual effects supervisors now report that bringing even the most challenging visions of film directors to the screen is no longer a question of whatDs possible; with todayDs techniques it is only a matter of time and cost. Driving this increase in realism have been computer graphics (CG) techniques for simulating how light travels within a scene and for simulating how light reflects off of and through surfaces. These techniquesJsome developed recently, and some originating in the 1980DsJare being applied to the visual effects process by computer graphics artists who have found ways to channel the power of these new tools.},
keywords = {Graphics},
pubstate = {published},
tppubtype = {inproceedings}
}
Anyone who has seen a recent summer blockbuster has witnessed the dramatic increases in computer-generated realism in recent years. Visual effects supervisors now report that bringing even the most challenging visions of film directors to the screen is no longer a question of whatDs possible; with todayDs techniques it is only a matter of time and cost. Driving this increase in realism have been computer graphics (CG) techniques for simulating how light travels within a scene and for simulating how light reflects off of and through surfaces. These techniquesJsome developed recently, and some originating in the 1980DsJare being applied to the visual effects process by computer graphics artists who have found ways to channel the power of these new tools.
Kallman, Marcelo; Marsella, Stacy C.
Hierarchical Motion Controllers for Real-Time Autonomous Virtual Humans Proceedings Article
In: International Working Conference on Intelligent Virtual Agents, Kos, Greece, 2005.
Abstract | Links | BibTeX | Tags: Virtual Humans
@inproceedings{kallman_hierarchical_2005,
title = {Hierarchical Motion Controllers for Real-Time Autonomous Virtual Humans},
author = {Marcelo Kallman and Stacy C. Marsella},
url = {http://ict.usc.edu/pubs/Hierarchical%20Motion%20Controllers%20for%20Real-Time%20Autonomous%20Virtual%20Humans.pdf},
year = {2005},
date = {2005-09-01},
booktitle = {International Working Conference on Intelligent Virtual Agents},
address = {Kos, Greece},
abstract = {Continuous and synchronized whole-body motions are essential for achieving believable autonomous virtual humans in interactive applications. We present a new motion control architecture based on generic controllers that can be hierarchically interconnected and reused in real-time. The hierarchical organization implies that leaf controllers are motion generators while the other nodes are connectors, performing operations such as interpolation, blending, and precise scheduling of children controllers. We also describe how the system can correctly handle the synchronization of gestures with speech in order to achieve believable conversational characters. For that purpose, different types of controllers implement a generic model of the different phases of a gesture.},
keywords = {Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
Continuous and synchronized whole-body motions are essential for achieving believable autonomous virtual humans in interactive applications. We present a new motion control architecture based on generic controllers that can be hierarchically interconnected and reused in real-time. The hierarchical organization implies that leaf controllers are motion generators while the other nodes are connectors, performing operations such as interpolation, blending, and precise scheduling of children controllers. We also describe how the system can correctly handle the synchronization of gestures with speech in order to achieve believable conversational characters. For that purpose, different types of controllers implement a generic model of the different phases of a gesture.
Kwon, Soon-il; Narayanan, Shrikanth
Unsupervised Speaker Indexing Using Generic Models Journal Article
In: IEEE Transactions on Speech and Audio Processing, vol. 13, no. 5, pp. 1004–1013, 2005.
Abstract | Links | BibTeX | Tags:
@article{kwon_unsupervised_2005,
title = {Unsupervised Speaker Indexing Using Generic Models},
author = {Soon-il Kwon and Shrikanth Narayanan},
url = {http://ict.usc.edu/pubs/Unsupervised%20Speaker%20Indexing%20Using%20Generic%20Models.pdf},
year = {2005},
date = {2005-09-01},
journal = {IEEE Transactions on Speech and Audio Processing},
volume = {13},
number = {5},
pages = {1004–1013},
abstract = {Unsupervised speaker indexing sequentially detects points where a speaker identity changes in a multispeaker audio stream, and categorizes each speaker segment, without any prior knowledge about the speakers. This paper addresses two chal- lenges: The first relates to sequential speaker change detection. The second relates to speaker modeling in light of the fact that the number/identity of the speakers is unknown. To address this issue, a predetermined generic speaker-independent model set, called the sample speaker models (SSM), is proposed. This set can be useful for more accurate speaker modeling and clustering without requiring training models on target speaker data. Once a speaker-independent model is selected from the generic sample models, it is progressively adapted into a specific speaker-depen- dent model. Experiments were performed with data from the Speaker Recognition Benchmark NIST Speech corpus (1999) and the HUB-4 Broadcast News Evaluation English Test material (1999). Results showed that our new technique, sampled using the Markov Chain Monte Carlo method, gave 92.5% indexing accuracy on two speaker telephone conversations, 89.6% on four-speaker conversations with the telephone speech quality, and 87.2% on broadcast news. The SSMs outperformed the universal background model by up to 29.4% and the universal gender models by up to 22.5% in indexing accuracy in the experiments of this paper.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Unsupervised speaker indexing sequentially detects points where a speaker identity changes in a multispeaker audio stream, and categorizes each speaker segment, without any prior knowledge about the speakers. This paper addresses two chal- lenges: The first relates to sequential speaker change detection. The second relates to speaker modeling in light of the fact that the number/identity of the speakers is unknown. To address this issue, a predetermined generic speaker-independent model set, called the sample speaker models (SSM), is proposed. This set can be useful for more accurate speaker modeling and clustering without requiring training models on target speaker data. Once a speaker-independent model is selected from the generic sample models, it is progressively adapted into a specific speaker-depen- dent model. Experiments were performed with data from the Speaker Recognition Benchmark NIST Speech corpus (1999) and the HUB-4 Broadcast News Evaluation English Test material (1999). Results showed that our new technique, sampled using the Markov Chain Monte Carlo method, gave 92.5% indexing accuracy on two speaker telephone conversations, 89.6% on four-speaker conversations with the telephone speech quality, and 87.2% on broadcast news. The SSMs outperformed the universal background model by up to 29.4% and the universal gender models by up to 22.5% in indexing accuracy in the experiments of this paper.
Jones, Andrew; Gardner, Andrew; Bolas, Mark; McDowall, Ian; Debevec, Paul
Performance Geometry Capture for Spatially Varying Relighting Proceedings Article
In: SIGGRAPH 2005 Sketch, Los Angeles, CA, 2005.
Links | BibTeX | Tags: Graphics, MxR
@inproceedings{jones_performance_2005,
title = {Performance Geometry Capture for Spatially Varying Relighting},
author = {Andrew Jones and Andrew Gardner and Mark Bolas and Ian McDowall and Paul Debevec},
url = {http://ict.usc.edu/pubs/Performance%20Geometry%20Capture%20for%20Spatially%20Varying%20Relighting.pdf},
year = {2005},
date = {2005-08-01},
booktitle = {SIGGRAPH 2005 Sketch},
address = {Los Angeles, CA},
keywords = {Graphics, MxR},
pubstate = {published},
tppubtype = {inproceedings}
}
Patel, Sanjit; Chu, Anson; Cohen, Jonathan; Pighin, Frédéric
Fluid Simulation Via Disjoint Translating Grids Proceedings Article
In: Special Interest Group - Graphics Technical Sketch, Los Angeles, CA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{patel_fluid_2005,
title = {Fluid Simulation Via Disjoint Translating Grids},
author = {Sanjit Patel and Anson Chu and Jonathan Cohen and Frédéric Pighin},
url = {http://ict.usc.edu/pubs/Fluid%20Simulation%20Via%20Disjoint%20Translating%20Grids.pdf},
year = {2005},
date = {2005-08-01},
booktitle = {Special Interest Group - Graphics Technical Sketch},
address = {Los Angeles, CA},
abstract = {We present an adaptive fluid simulation technique that splits the computation domain in multiple moving grids. Using this technique, we are able to simulate fluids over large spatial domains with reasonable computation times.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present an adaptive fluid simulation technique that splits the computation domain in multiple moving grids. Using this technique, we are able to simulate fluids over large spatial domains with reasonable computation times.
Debevec, Paul
A Median Cut Algorithm for Light Probe Sampling Proceedings Article
In: SIGGRAPH (Special Interest Group - Graphics), Los Angeles, CA, 2005.
Abstract | Links | BibTeX | Tags: Graphics
@inproceedings{debevec_median_2005,
title = {A Median Cut Algorithm for Light Probe Sampling},
author = {Paul Debevec},
url = {http://ict.usc.edu/pubs/A%20Median%20Cut%20Algorithm%20for%20Light%20Probe%20Sampling.pdf},
year = {2005},
date = {2005-08-01},
booktitle = {SIGGRAPH (Special Interest Group - Graphics)},
address = {Los Angeles, CA},
abstract = {We present a technique for approximating a light probe image as a constellation of light sources based on a median cut algorithm. The algorithm is efï¬cient, simple to implement, and can realistically represent a complex lighting environment with as few as 64 point light sources.},
keywords = {Graphics},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a technique for approximating a light probe image as a constellation of light sources based on a median cut algorithm. The algorithm is efï¬cient, simple to implement, and can realistically represent a complex lighting environment with as few as 64 point light sources.
Morie, Jacquelyn; Williams, Josh; Dozois, Aimee; Luigi, Donat-Pierre
The Fidelity of "Feel": Emotional Affordance in Virtual Environments Proceedings Article
In: 11th International Conference on Human-Computer Interaction, Las Vegas, NV, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{morie_fidelity_2005,
title = {The Fidelity of "Feel": Emotional Affordance in Virtual Environments},
author = {Jacquelyn Morie and Josh Williams and Aimee Dozois and Donat-Pierre Luigi},
url = {http://ict.usc.edu/pubs/The%20Fidelity%20of%20Feel-%20Emotional%20Affordance%20in%20Virtual%20Environments.pdf},
year = {2005},
date = {2005-07-01},
booktitle = {11th International Conference on Human-Computer Interaction},
address = {Las Vegas, NV},
abstract = {Virtual environments (VEs) should be able to provide experiences as rich and complex as those to be had in real life. While this seems obvious, it is not yet possible to create a perfect simulacrum of the real world, so such correspondence requires the development of design techniques by which VEs can be made to appear more real. It also requires evaluation studies to determine if such techniques produce the desired results. As emotions are implicated in our phenomenological understanding of the physical world, they should also play an integral role in the experience of the virtual one. Therefore, a logical sequence of experimentation to understand how VEs can be made to function as emotion-induction systems is in order. The Sensory Environments Evaluation (SEE) research program has developed a twofold design process to explore if we react to virtually supplied stimuli as we do to the real world equivalents. We look at manipulating both the sensory and emotional aspects of not only the environment but also the participant. We do this with the focus on what emotional affordances this manipulation will provide. Our first evaluation scenario, DarkCon, was designed in this way to produce a strong sense of presence. Sixty-four subjects have been fielded to date and the data is currently being analyzed for results. We hope to find that rich design techniques along with the frame of mind with which a VR experience is approached will predictably influence perception and behavior within a virtual world. We will use these results to inform continuing research into the creation of more emotionally affective VEs.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Virtual environments (VEs) should be able to provide experiences as rich and complex as those to be had in real life. While this seems obvious, it is not yet possible to create a perfect simulacrum of the real world, so such correspondence requires the development of design techniques by which VEs can be made to appear more real. It also requires evaluation studies to determine if such techniques produce the desired results. As emotions are implicated in our phenomenological understanding of the physical world, they should also play an integral role in the experience of the virtual one. Therefore, a logical sequence of experimentation to understand how VEs can be made to function as emotion-induction systems is in order. The Sensory Environments Evaluation (SEE) research program has developed a twofold design process to explore if we react to virtually supplied stimuli as we do to the real world equivalents. We look at manipulating both the sensory and emotional aspects of not only the environment but also the participant. We do this with the focus on what emotional affordances this manipulation will provide. Our first evaluation scenario, DarkCon, was designed in this way to produce a strong sense of presence. Sixty-four subjects have been fielded to date and the data is currently being analyzed for results. We hope to find that rich design techniques along with the frame of mind with which a VR experience is approached will predictably influence perception and behavior within a virtual world. We will use these results to inform continuing research into the creation of more emotionally affective VEs.
Lane, H. Chad; Core, Mark; Lent, Michael; Solomon, Steve; Gomboc, Dave
Explainable Artificial Intelligence for Training and Tutoring Proceedings Article
In: 12th International Conference on Artificial Intelligence in Education, Amsterdam, The Netherlands, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{lane_explainable_2005,
title = {Explainable Artificial Intelligence for Training and Tutoring},
author = {H. Chad Lane and Mark Core and Michael Lent and Steve Solomon and Dave Gomboc},
url = {http://ict.usc.edu/pubs/Explainable%20Artificial%20Intelligence%20for%20Training%20and%20Tutoring.pdf},
year = {2005},
date = {2005-07-01},
booktitle = {12th International Conference on Artificial Intelligence in Education},
address = {Amsterdam, The Netherlands},
abstract = {This paper describes an Explainable Artificial Intelligence (XAI) tool that allows entities to answer questions about their activities within a tactical simulation. We show how XAI can be used to provide more meaningful after-action reviews and discuss ongoing work to integrate an intelligent tutor into the XAI framework.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper describes an Explainable Artificial Intelligence (XAI) tool that allows entities to answer questions about their activities within a tactical simulation. We show how XAI can be used to provide more meaningful after-action reviews and discuss ongoing work to integrate an intelligent tutor into the XAI framework.
Traum, David; Swartout, William; Gratch, Jonathan; Marsella, Stacy C.
Virtual Humans for non-team interaction training Proceedings Article
In: International Conference on Autonomous Agents and Multiagent Systems (AAMAS) Workshop on Creating Bonds with Humanoids, Utrecht, Netherlands, 2005.
Abstract | Links | BibTeX | Tags: Social Simulation, Virtual Humans
@inproceedings{traum_virtual_2005,
title = {Virtual Humans for non-team interaction training},
author = {David Traum and William Swartout and Jonathan Gratch and Stacy C. Marsella},
url = {http://ict.usc.edu/pubs/Virtual%20Humans%20for%20non-team%20interaction%20training.pdf},
year = {2005},
date = {2005-07-01},
booktitle = {International Conference on Autonomous Agents and Multiagent Systems (AAMAS) Workshop on Creating Bonds with Humanoids},
address = {Utrecht, Netherlands},
abstract = {We describe a model of virtual humans to be used in training for non-team interactions, such as negotiating with people from other organizations. The virtual humans build on existing task, dialogue, and emotion models, with an added model of trust, which are used to understand and produce interactional moves. The model has been implemented within an agent in the SASO-ST system, and some example dialogues are given, illustrating the necessity for building social bonds.},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
We describe a model of virtual humans to be used in training for non-team interactions, such as negotiating with people from other organizations. The virtual humans build on existing task, dialogue, and emotion models, with an added model of trust, which are used to understand and produce interactional moves. The model has been implemented within an agent in the SASO-ST system, and some example dialogues are given, illustrating the necessity for building social bonds.
Gratch, Jonathan; Marsella, Stacy C.
Evaluating a computational model of emotion Journal Article
In: Journal Autonomous Agents and Multi-Agent Systems. Special Issue on the Best of AAMAS 2004, vol. 11, no. 1, pp. 23–43, 2005.
Abstract | Links | BibTeX | Tags: Social Simulation, Virtual Humans
@article{gratch_evaluating_2005,
title = {Evaluating a computational model of emotion},
author = {Jonathan Gratch and Stacy C. Marsella},
url = {http://ict.usc.edu/pubs/Evaluating%20a%20computational%20model%20of%20emotion.pdf},
year = {2005},
date = {2005-07-01},
journal = {Journal Autonomous Agents and Multi-Agent Systems. Special Issue on the Best of AAMAS 2004},
volume = {11},
number = {1},
pages = {23–43},
abstract = {Spurred by a range of potential applications, there has been a growing body of research in computational models of human emotion. To advance the development of these models, it is critical that we evaluate them against the phenomena they purport to model. In this paper, we present one method to evaluate an emotion model that compares the behavior of the model against human behavior using a standard clinical instrument for assessing human emotion and coping. We use this method to evaluate the Emotion and Adaptation (EMA) model of emotion Gratch and Marsella. The evaluation highlights strengths of the approach and identifies where the model needs further development.},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {article}
}
Spurred by a range of potential applications, there has been a growing body of research in computational models of human emotion. To advance the development of these models, it is critical that we evaluate them against the phenomena they purport to model. In this paper, we present one method to evaluate an emotion model that compares the behavior of the model against human behavior using a standard clinical instrument for assessing human emotion and coping. We use this method to evaluate the Emotion and Adaptation (EMA) model of emotion Gratch and Marsella. The evaluation highlights strengths of the approach and identifies where the model needs further development.
Riedl, Mark O.; Lane, H. Chad; Hill, Randall W.; Swartout, William
Automated Story Direction and Intelligent Tutoring: Towards a Unifying Architecture Proceedings Article
In: AI and Education 2005 Workshop on Narrative Learning Environments, Amsterdam, The Netherlands, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{riedl_automated_2005,
title = {Automated Story Direction and Intelligent Tutoring: Towards a Unifying Architecture},
author = {Mark O. Riedl and H. Chad Lane and Randall W. Hill and William Swartout},
url = {http://ict.usc.edu/pubs/Automated%20Story%20Direction%20and%20Intelligent%20Tutoring-%20Towards%20a%20Unifying%20Architecture.pdf},
year = {2005},
date = {2005-07-01},
booktitle = {AI and Education 2005 Workshop on Narrative Learning Environments},
address = {Amsterdam, The Netherlands},
abstract = {Recently, interactive storytelling systems H systems that allow a user to make decisions that can potentially impact the direction of a narrative H have been applied to training and education. Interactive storytelling systems often rely on an automated story director to manage the userKs experience. The focus of an automated director is the emergence of a narrative-like experience for the user. In contrast, intelligent tutors traditionally address the acquisition or strengthening of a learner's knowledge. Our goal is to build training simulations that cultivate compelling storylines while simultaneously maintaining a pedagogical presence by incorporating both automated story direction and intelligent tutoring into an immersive environment. But what is the relationship between an automated director and an intelligent tutor? In this paper, we discuss the similarities and differences of automated story directors and intelligent tutors and, based on our analysis, recommend an architecture for building narrative-based training simulations that utilize both effectively and without conflict.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Recently, interactive storytelling systems H systems that allow a user to make decisions that can potentially impact the direction of a narrative H have been applied to training and education. Interactive storytelling systems often rely on an automated story director to manage the userKs experience. The focus of an automated director is the emergence of a narrative-like experience for the user. In contrast, intelligent tutors traditionally address the acquisition or strengthening of a learner's knowledge. Our goal is to build training simulations that cultivate compelling storylines while simultaneously maintaining a pedagogical presence by incorporating both automated story direction and intelligent tutoring into an immersive environment. But what is the relationship between an automated director and an intelligent tutor? In this paper, we discuss the similarities and differences of automated story directors and intelligent tutors and, based on our analysis, recommend an architecture for building narrative-based training simulations that utilize both effectively and without conflict.
Rizzo, Albert; Kim, Gerard J.; Yeh, Shih-Ching; Thiebaux, Marcus; Hwang, Jayne; Buckwalter, John Galen
Development of a Benchmarking Scenario for Testing 3D User Interface Devices and Interaction Methods Proceedings Article
In: Proceedings of the 11th International Conference on Human Computer Interaction, Las Vegas, NV, 2005.
Abstract | Links | BibTeX | Tags: MedVR
@inproceedings{rizzo_development_2005,
title = {Development of a Benchmarking Scenario for Testing 3D User Interface Devices and Interaction Methods},
author = {Albert Rizzo and Gerard J. Kim and Shih-Ching Yeh and Marcus Thiebaux and Jayne Hwang and John Galen Buckwalter},
url = {http://ict.usc.edu/pubs/Development%20of%20a%20Benchmarking%20Scenario%20for%20Testing%203D%20User%20Interface%20Devices%20and%20Interaction%20Methods.pdf},
year = {2005},
date = {2005-07-01},
booktitle = {Proceedings of the 11th International Conference on Human Computer Interaction},
address = {Las Vegas, NV},
abstract = {To address a part of the challenge of testing and comparing various 3D user interface devices and methods, we are currently developing and testing a VR 3D User Interface benchmarking scenario. The approach outlined in this paper focuses on the capture of human interaction performance on object selection and manipulation tasks using standardized and scalable block configurations that allow for measurement of speed and efficiency with any interaction device or method. The block configurations that we are using as benchmarking stimuli are accompanied by a pure mental rotation visuospatial assessment test. This feature will allow researchers to test usersX existing spatial abilities and statistically parcel out the variability due to innate ability, from the actual hands-on performance metrics. This statistical approach could lead to a more pure analysis of the ergonomic features of interaction devices and methods separate from existing user abilities. An initial test was conducted at two sites using this benchmarking system to make comparisons between 3D/gesture-based and 2D/mouse-based interactions for 3D selection and manipulation. Our preliminary results demonstrated, as expected, that the 3D/gesture based method in general outperformed the 2D/mouse interface. As well there were statistically significant performance differences between different user groups when categorized by their sex, visuospatial ability and educational background.},
keywords = {MedVR},
pubstate = {published},
tppubtype = {inproceedings}
}
To address a part of the challenge of testing and comparing various 3D user interface devices and methods, we are currently developing and testing a VR 3D User Interface benchmarking scenario. The approach outlined in this paper focuses on the capture of human interaction performance on object selection and manipulation tasks using standardized and scalable block configurations that allow for measurement of speed and efficiency with any interaction device or method. The block configurations that we are using as benchmarking stimuli are accompanied by a pure mental rotation visuospatial assessment test. This feature will allow researchers to test usersX existing spatial abilities and statistically parcel out the variability due to innate ability, from the actual hands-on performance metrics. This statistical approach could lead to a more pure analysis of the ergonomic features of interaction devices and methods separate from existing user abilities. An initial test was conducted at two sites using this benchmarking system to make comparisons between 3D/gesture-based and 2D/mouse-based interactions for 3D selection and manipulation. Our preliminary results demonstrated, as expected, that the 3D/gesture based method in general outperformed the 2D/mouse interface. As well there were statistically significant performance differences between different user groups when categorized by their sex, visuospatial ability and educational background.
Gordon, Andrew S.
Commonsense Psychology and the Functional Requirements of Cognitive Models Proceedings Article
In: American Association of Artificial Intelligence Workshop on Modular Construction of Human-Like Intelligence, AAAI Press, Pittsburgh, PA, 2005.
Abstract | Links | BibTeX | Tags: The Narrative Group
@inproceedings{gordon_commonsense_2005,
title = {Commonsense Psychology and the Functional Requirements of Cognitive Models},
author = {Andrew S. Gordon},
url = {http://ict.usc.edu/pubs/Commonsense%20Psychology%20and%20the%20Functional%20Requirements%20of%20Cognitive%20Models.pdf},
year = {2005},
date = {2005-07-01},
booktitle = {American Association of Artificial Intelligence Workshop on Modular Construction of Human-Like Intelligence},
publisher = {AAAI Press},
address = {Pittsburgh, PA},
abstract = {In this paper we argue that previous models of cognitive abilities (e.g. memory, analogy) have been constructed to satisfy functional requirements of implicit commonsense psychological theories held by researchers and nonresearchers alike. Rather than working to avoid the influence of commonsense psychology in cognitive modeling research, we propose to capitalize on progress in developing formal theories of commonsense psychology to explicitly define the functional requirements of cognitive models. We present a taxonomy of 16 classes of cognitive models that correspond to the representational areas that have been addressed in large-scale inferential theories of commonsense psychology. We consider the functional requirements that can be derived from inferential theories for one of these classes, the processes involved in human memory. We argue that the breadth coverage of commonsense theories can be used to better evaluate the explanatory scope of cognitive models, as well as facilitate the investigation of larger-scale cognitive systems.},
keywords = {The Narrative Group},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper we argue that previous models of cognitive abilities (e.g. memory, analogy) have been constructed to satisfy functional requirements of implicit commonsense psychological theories held by researchers and nonresearchers alike. Rather than working to avoid the influence of commonsense psychology in cognitive modeling research, we propose to capitalize on progress in developing formal theories of commonsense psychology to explicitly define the functional requirements of cognitive models. We present a taxonomy of 16 classes of cognitive models that correspond to the representational areas that have been addressed in large-scale inferential theories of commonsense psychology. We consider the functional requirements that can be derived from inferential theories for one of these classes, the processes involved in human memory. We argue that the breadth coverage of commonsense theories can be used to better evaluate the explanatory scope of cognitive models, as well as facilitate the investigation of larger-scale cognitive systems.
Ettaile, Emil; Gandhe, Sudeep; Georgiou, Panayiotis G.; Knight, Kevin; Marcu, Daniel; Narayanan, Shrikanth; Traum, David; Belvin, Robert
Transonics: A Practical Speech-to-Speech Translator for English-Farsi Medical Dialogues Proceedings Article
In: Proceedings of the ACL Interactive Poster and Demonstration Sessions, pp. 89–92, Ann Arbor, MI, 2005.
Abstract | Links | BibTeX | Tags: Virtual Humans
@inproceedings{ettaile_transonics_2005,
title = {Transonics: A Practical Speech-to-Speech Translator for English-Farsi Medical Dialogues},
author = {Emil Ettaile and Sudeep Gandhe and Panayiotis G. Georgiou and Kevin Knight and Daniel Marcu and Shrikanth Narayanan and David Traum and Robert Belvin},
url = {http://ict.usc.edu/pubs/TRANSONICS-%20A%20SPEECH%20TO%20SPEECH%20SYSTEM%20FOR%20ENGLISH-PERSIAN%20INTERACTIONS.pdf},
year = {2005},
date = {2005-06-01},
booktitle = {Proceedings of the ACL Interactive Poster and Demonstration Sessions},
pages = {89–92},
address = {Ann Arbor, MI},
abstract = {We briefly describe a two-way speech-to-speech English-Farsi translation system prototype developed for use in doctorpatient interactions. The overarching philosophy of the developers has been to create a system that enables effective communication, rather than focusing on maximizing component-level performance. The discussion focuses on the general approach and evaluation of the system by an independent government evaluation team.},
keywords = {Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
We briefly describe a two-way speech-to-speech English-Farsi translation system prototype developed for use in doctorpatient interactions. The overarching philosophy of the developers has been to create a system that enables effective communication, rather than focusing on maximizing component-level performance. The discussion focuses on the general approach and evaluation of the system by an independent government evaluation team.
Liao, Wei-Kai; Cohen, Isaac
Classifying Facial Gestures in Presence of Head Motion Proceedings Article
In: IEEE Workshop on Vision for Human-Computer Interaction, San Diego, CA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{liao_classifying_2005,
title = {Classifying Facial Gestures in Presence of Head Motion},
author = {Wei-Kai Liao and Isaac Cohen},
url = {http://ict.usc.edu/pubs/Classifying%20Facial%20Gestures%20in%20Presence%20of%20Head%20Motion.pdf},
year = {2005},
date = {2005-06-01},
booktitle = {IEEE Workshop on Vision for Human-Computer Interaction},
address = {San Diego, CA},
abstract = {This paper addresses the problem of automatic facial gestures recognition in an interactive environment. Automatic facial gestures recognition is a difficult problem in computer vision, and most of the work has focused on inferring facial gestures in the context of a static head. In the paper we address the challenging problem of recognizing the facial expressions of a moving head. We present a systematic framework to analyze and classify the facial gestures with the head movement. Our system includes a 3D head pose estimation method to recover the global head motion. After estimating the head pose, the human face is modeled by a collection of face's regions. These regions represent the face model used for locating and extracting temporal facial features. We propose using a locally affine motion model to represent extracted motion fields. The classification consists of a graphical model for robustly representing the dependencies of the selected facial regions and the support vector machine. Our experiments show that this approach could classify human expressions in interactive environments accurately.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper addresses the problem of automatic facial gestures recognition in an interactive environment. Automatic facial gestures recognition is a difficult problem in computer vision, and most of the work has focused on inferring facial gestures in the context of a static head. In the paper we address the challenging problem of recognizing the facial expressions of a moving head. We present a systematic framework to analyze and classify the facial gestures with the head movement. Our system includes a 3D head pose estimation method to recover the global head motion. After estimating the head pose, the human face is modeled by a collection of face's regions. These regions represent the face model used for locating and extracting temporal facial features. We propose using a locally affine motion model to represent extracted motion fields. The classification consists of a graphical model for robustly representing the dependencies of the selected facial regions and the support vector machine. Our experiments show that this approach could classify human expressions in interactive environments accurately.
Chu, Chi-Wei; Cohen, Isaac
Posture and Gesture Recognition using 3D Body Shapes Decomposition Proceedings Article
In: IEEE Workshop on Vision for Human-Computer Interaction, San Diego, CA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{chu_posture_2005,
title = {Posture and Gesture Recognition using 3D Body Shapes Decomposition},
author = {Chi-Wei Chu and Isaac Cohen},
url = {http://ict.usc.edu/pubs/Posture%20and%20Gesture%20Recognition%20using%203D%20Body%20Shapes%20Decomposition.pdf},
year = {2005},
date = {2005-06-01},
booktitle = {IEEE Workshop on Vision for Human-Computer Interaction},
address = {San Diego, CA},
abstract = {We present a method for describing arbitrary human posture as a combination of basic postures. This decomposition allows for recognition of a larger number of postures and gestures from a small set of elementary postures called atoms. We propose a modified version of the matching pursuit algorithm for decomposing an arbitrary input posture into a linear combination of primary and secondary atoms. These atoms are represented through their shape descriptor inferred from the 3D visual-hull of the human body posture. Using an atom-based description of postures increases tremendously the set of recognizable postures while reducing the required training data set. A gesture recognition system based on the atom decomposition and Hidden Markov Model (HMM) is also described. Instead of representing gestures as HMM transition of postures, we separate the description of gestures as two HMMs, each describing the transition of Primary/Secondary atoms; thus greatly reducing the size of state space of HMM. We illustrate the proposed approach for posture and gesture recognition method on a set of video streams captured by four synchronous cameras.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a method for describing arbitrary human posture as a combination of basic postures. This decomposition allows for recognition of a larger number of postures and gestures from a small set of elementary postures called atoms. We propose a modified version of the matching pursuit algorithm for decomposing an arbitrary input posture into a linear combination of primary and secondary atoms. These atoms are represented through their shape descriptor inferred from the 3D visual-hull of the human body posture. Using an atom-based description of postures increases tremendously the set of recognizable postures while reducing the required training data set. A gesture recognition system based on the atom decomposition and Hidden Markov Model (HMM) is also described. Instead of representing gestures as HMM transition of postures, we separate the description of gestures as two HMMs, each describing the transition of Primary/Secondary atoms; thus greatly reducing the size of state space of HMM. We illustrate the proposed approach for posture and gesture recognition method on a set of video streams captured by four synchronous cameras.
Kim, Youngjun; Velson, Martin; Hill, Randall W.
Modeling Dynamic Perceptual Attention in Complex Virtual Environments Proceedings Article
In: Conference on Behavior Representation in Modeling and Simulation (BRIMS), Universal City, CA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{kim_modeling_2005,
title = {Modeling Dynamic Perceptual Attention in Complex Virtual Environments},
author = {Youngjun Kim and Martin Velson and Randall W. Hill},
url = {http://ict.usc.edu/pubs/Modeling%20Dynamic%20Perceptual%20Attention%20in%20Complex%20Virtual%20Environments.pdf},
year = {2005},
date = {2005-05-01},
booktitle = {Conference on Behavior Representation in Modeling and Simulation (BRIMS)},
address = {Universal City, CA},
abstract = {An important characteristic of a virtual human is the ability to direct its perceptual attention to entities and areas in a virtual environment in a manner that appears believable and serves a functional purpose. In this paper, we describe a perceptual attention model that integrates perceptual attention that mediates top-down and bottom-up attention processes of virtual humans within complex virtual environments.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
An important characteristic of a virtual human is the ability to direct its perceptual attention to entities and areas in a virtual environment in a manner that appears believable and serves a functional purpose. In this paper, we describe a perceptual attention model that integrates perceptual attention that mediates top-down and bottom-up attention processes of virtual humans within complex virtual environments.
Hobbs, Jerry R.; Gordon, Andrew S.
Encoding Knowledge of Commonsense Psychology Proceedings Article
In: 7th International Symposium on Logical Formalizations of Commonsense Reasoning, Corfu, Greece, 2005.
Abstract | Links | BibTeX | Tags: The Narrative Group
@inproceedings{hobbs_encoding_2005,
title = {Encoding Knowledge of Commonsense Psychology},
author = {Jerry R. Hobbs and Andrew S. Gordon},
url = {http://ict.usc.edu/pubs/Encoding%20Knowledge%20of%20Commonsense%20Psychology.pdf},
year = {2005},
date = {2005-05-01},
booktitle = {7th International Symposium on Logical Formalizations of Commonsense Reasoning},
address = {Corfu, Greece},
abstract = {Introduction: In previous papers (Gordon and Hobbs, 2003, 2004) we have described a methodology for determining what knowledge should be included in the knowledge base for an intelligent agent, capable of constructing and executing plans to achieve its goals. An intelligent agent is at least a planning mechanism, so Gordon (2004) asked what concepts are necessary for the common strategies that people use in achieving their goals. He investigated ten different domains, including politics, personal relationships, artistic performance, and warfare, and collected 372 strategies. He authored representations of these strategies in order to identify a controlled vocabulary involving of concepts. These concepts were categorized into 48 different representational areas, such as sets, space, and time. Thirty of the representational areas, involving 635 concepts, were concerned with commonsense psychology; among these are memory, knowledge management, planning, and so on. This result by itself demonstrates the very great importance of commonsense psychology in the construction of intelligent agents. Gordon et al. (2003) then, to deï¬ne further each of the representational areas, augmented the list of concepts by investigating the English language expressions for concepts in each area. The result was a list of 528 concepts, a set that identiï¬es the target coverage of a formal theory of commonsense psychology. The authors began the development of formal theories that would encompass this list of concepts. In our earlier work (Gordon and Hobbs, 2003), we described the ï¬rst theory we constructed, memory, as an illustration of the method. We have now completed 14 of the 30 theories, and this paper provides an overview of this work as we close in on the halfway mark.},
keywords = {The Narrative Group},
pubstate = {published},
tppubtype = {inproceedings}
}
Introduction: In previous papers (Gordon and Hobbs, 2003, 2004) we have described a methodology for determining what knowledge should be included in the knowledge base for an intelligent agent, capable of constructing and executing plans to achieve its goals. An intelligent agent is at least a planning mechanism, so Gordon (2004) asked what concepts are necessary for the common strategies that people use in achieving their goals. He investigated ten different domains, including politics, personal relationships, artistic performance, and warfare, and collected 372 strategies. He authored representations of these strategies in order to identify a controlled vocabulary involving of concepts. These concepts were categorized into 48 different representational areas, such as sets, space, and time. Thirty of the representational areas, involving 635 concepts, were concerned with commonsense psychology; among these are memory, knowledge management, planning, and so on. This result by itself demonstrates the very great importance of commonsense psychology in the construction of intelligent agents. Gordon et al. (2003) then, to deï¬ne further each of the representational areas, augmented the list of concepts by investigating the English language expressions for concepts in each area. The result was a list of 528 concepts, a set that identiï¬es the target coverage of a formal theory of commonsense psychology. The authors began the development of formal theories that would encompass this list of concepts. In our earlier work (Gordon and Hobbs, 2003), we described the ï¬rst theory we constructed, memory, as an illustration of the method. We have now completed 14 of the 30 theories, and this paper provides an overview of this work as we close in on the halfway mark.
Kim, Youngjun; Hill, Randall W.; Traum, David
A Computational Model of Dynamic Perceptual Attention for Virtual Humans Proceedings Article
In: Proceedings of the 14th Conference on Behavior Representation in Modeling and Simulation, Universal City, CA, 2005.
Abstract | Links | BibTeX | Tags: Virtual Humans
@inproceedings{kim_computational_2005,
title = {A Computational Model of Dynamic Perceptual Attention for Virtual Humans},
author = {Youngjun Kim and Randall W. Hill and David Traum},
url = {http://ict.usc.edu/pubs/A%20Computational%20Model%20of%20Dynamic%20Perceptual%20Attention%20for%20Virtual%20Humans.pdf},
year = {2005},
date = {2005-05-01},
booktitle = {Proceedings of the 14th Conference on Behavior Representation in Modeling and Simulation},
address = {Universal City, CA},
abstract = {An important characteristic of a virtual human is the ability to direct its perceptual attention to objects and locations in a virtual environment in a manner that looks believable and serves a functional purpose. We have developed a computational model of perceptual attention that mediates top-down and bottom-up attention processes of virtual humans in virtual environments. In this paper, we propose a perceptual attention model that will integrate perceptual attention toward objects and locations in the environment with the need to look at other parties in a social context.},
keywords = {Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
An important characteristic of a virtual human is the ability to direct its perceptual attention to objects and locations in a virtual environment in a manner that looks believable and serves a functional purpose. We have developed a computational model of perceptual attention that mediates top-down and bottom-up attention processes of virtual humans in virtual environments. In this paper, we propose a perceptual attention model that will integrate perceptual attention toward objects and locations in the environment with the need to look at other parties in a social context.
Gomboc, Dave; Solomon, Steve; Core, Mark; Lane, H. Chad; Lent, Michael
Design Recommendations to Support Automated Explanation and Tutoring Proceedings Article
In: Conference on Behavior Representation in Modeling and Simulation (BRIMS), Universal City, CA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{gomboc_design_2005,
title = {Design Recommendations to Support Automated Explanation and Tutoring},
author = {Dave Gomboc and Steve Solomon and Mark Core and H. Chad Lane and Michael Lent},
url = {http://ict.usc.edu/pubs/Design%20Recommendations%20to%20Support%20Automated%20Explanation%20and%20Tutoring.pdf},
year = {2005},
date = {2005-05-01},
booktitle = {Conference on Behavior Representation in Modeling and Simulation (BRIMS)},
address = {Universal City, CA},
abstract = {The after-action review is an essential component of military training exercises. The use of constructive simulations for training poses a challenge when conducting such reviews, because behavior models are typically designed to simulate satisfactorially, without explicit concern for the interrogation of synthetic entities afterward. Ideally, users could obtain knowledge about not only the choices made by a simulatorEs behavior models, but also the rationale for those choices. This requires a rich representation of behavioral knowledge within the software system. We have integrated our explainable AI system with behavior models and log information from two simulation systems. Selecting examples from these simulators, we identify areas for improvement to facilitate the automation of explanation and tutoring.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The after-action review is an essential component of military training exercises. The use of constructive simulations for training poses a challenge when conducting such reviews, because behavior models are typically designed to simulate satisfactorially, without explicit concern for the interrogation of synthetic entities afterward. Ideally, users could obtain knowledge about not only the choices made by a simulatorEs behavior models, but also the rationale for those choices. This requires a rich representation of behavioral knowledge within the software system. We have integrated our explainable AI system with behavior models and log information from two simulation systems. Selecting examples from these simulators, we identify areas for improvement to facilitate the automation of explanation and tutoring.
Feng, Donghui; Hovy, Eduard
MRE: A Study on Evolutionary Language Understanding Proceedings Article
In: Second International Workshop on Natural Language Understanding and Cognitive Science (NLUCS), Miami, Florida, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{feng_mre_2005,
title = {MRE: A Study on Evolutionary Language Understanding},
author = {Donghui Feng and Eduard Hovy},
url = {http://ict.usc.edu/pubs/MRE-%20A%20Study%20on%20Evolutionary%20Language%20Understanding.pdf},
year = {2005},
date = {2005-05-01},
booktitle = {Second International Workshop on Natural Language Understanding and Cognitive Science (NLUCS)},
address = {Miami, Florida},
abstract = {The lack of well-annotated data is always one of the biggest problems for most training-based dialogue systems. Without enough training data, it's almost impossible for a trainable system to work. In this paper, we explore the evolutionary language understanding approach to build a natural language understanding machine in a virtual human training project. We build the initial training data with a finite state machine. The language understanding system is trained based on the automated data first and is improved as more and more real data come in, which is proved by the experimental results.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The lack of well-annotated data is always one of the biggest problems for most training-based dialogue systems. Without enough training data, it's almost impossible for a trainable system to work. In this paper, we explore the evolutionary language understanding approach to build a natural language understanding machine in a virtual human training project. We build the initial training data with a finite state machine. The language understanding system is trained based on the automated data first and is improved as more and more real data come in, which is proved by the experimental results.
Nijholt, Anton; Traum, David
The Virtuality Continuum Revisited Proceedings Article
In: CHI 2005 Workshop on the Virtuality Continuum Revisited, Portland, OR, 2005.
Abstract | Links | BibTeX | Tags: Virtual Humans
@inproceedings{nijholt_virtuality_2005,
title = {The Virtuality Continuum Revisited},
author = {Anton Nijholt and David Traum},
url = {http://ict.usc.edu/pubs/The%20Virtuality%20Continuum%20Revisited.pdf},
year = {2005},
date = {2005-04-01},
booktitle = {CHI 2005 Workshop on the Virtuality Continuum Revisited},
address = {Portland, OR},
abstract = {We survey the themes and the aims of a workshop devoted to the state-of-the-art virtuality continuum. In this continuum, ranging from fully virtual to real physical environments, allowing for mixed, augmented and desktop virtual reality, several perspectives can be taken. Originally, the emphasis was on display technologies. Here we take the perspective of the inhabited environment, that is, environments positioned somewhere on this continuum that are inhabited by virtual (embodied) agents, that interact with each other and with their human partners. Hence, we look at it from the multi-party interaction perspective. In this workshop we will investigate the current state of the art, its shortcomings and a future research agenda.},
keywords = {Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
We survey the themes and the aims of a workshop devoted to the state-of-the-art virtuality continuum. In this continuum, ranging from fully virtual to real physical environments, allowing for mixed, augmented and desktop virtual reality, several perspectives can be taken. Originally, the emphasis was on display technologies. Here we take the perspective of the inhabited environment, that is, environments positioned somewhere on this continuum that are inhabited by virtual (embodied) agents, that interact with each other and with their human partners. Hence, we look at it from the multi-party interaction perspective. In this workshop we will investigate the current state of the art, its shortcomings and a future research agenda.
Peterson, Michael J.; Kyriakakis, Chris
Choosing Candidate Locations for Source Localization Proceedings Article
In: International Workshop on Hands Free Communication and Microphone Arrays, Rutgers, NY, 2005.
@inproceedings{peterson_choosing_2005,
title = {Choosing Candidate Locations for Source Localization},
author = {Michael J. Peterson and Chris Kyriakakis},
url = {http://ict.usc.edu/pubs/CHOOSING%20CANDIDATE%20LOCATIONS%20FOR%20SOURCE%20LOCALIZATION.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {International Workshop on Hands Free Communication and Microphone Arrays},
address = {Rutgers, NY},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Swanson, Reid; Gordon, Andrew S.
Automated Commonsense Reasoning About Human Memory Proceedings Article
In: AAAI Spring Symposium on Metacognitive Computing, Stanford, CA, 2005.
Abstract | Links | BibTeX | Tags: The Narrative Group
@inproceedings{swanson_automated_2005,
title = {Automated Commonsense Reasoning About Human Memory},
author = {Reid Swanson and Andrew S. Gordon},
url = {http://ict.usc.edu/pubs/Automated%20Commonsense%20Reasoning%20About%20Human%20Memory.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {AAAI Spring Symposium on Metacognitive Computing},
address = {Stanford, CA},
abstract = {Metacognitive reasoning in computational systems will be enabled by the development of formal theories that have broad coverage over mental states and processes as well as inferential competency. In this paper we evaluate the inferential competency of an existing formal theory of commonsense human memory by attempting to use it to validate the appropriateness of a commonsense memory strategy. We formulate a particular memory strategy (to create an associated obstacle) as a theorem in first-order predicate calculus. We then attempt to validate this strategy by showing that it is entailed by the axioms of the theory we evaluated. These axioms were encoded into the syntax of an automated reasoning system, which was used to automatically generate inferences and search for formal proofs.},
keywords = {The Narrative Group},
pubstate = {published},
tppubtype = {inproceedings}
}
Metacognitive reasoning in computational systems will be enabled by the development of formal theories that have broad coverage over mental states and processes as well as inferential competency. In this paper we evaluate the inferential competency of an existing formal theory of commonsense human memory by attempting to use it to validate the appropriateness of a commonsense memory strategy. We formulate a particular memory strategy (to create an associated obstacle) as a theorem in first-order predicate calculus. We then attempt to validate this strategy by showing that it is entailed by the axioms of the theory we evaluated. These axioms were encoded into the syntax of an automated reasoning system, which was used to automatically generate inferences and search for formal proofs.
Busso, Carlos; Hernanz, Sergi; Chu, Chi-Wei; Kwon, Soon-il; Lee, Sung; Georgiou, Panayiotis G.; Cohen, Isaac; Narayanan, Shrikanth
Smart Room: Participant and Speaker Localization and Identification Proceedings Article
In: Proceedings of the IEEE Conference on Acoustics, Speech and Signal Processing, Philadelphia, PA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{busso_smart_2005,
title = {Smart Room: Participant and Speaker Localization and Identification},
author = {Carlos Busso and Sergi Hernanz and Chi-Wei Chu and Soon-il Kwon and Sung Lee and Panayiotis G. Georgiou and Isaac Cohen and Shrikanth Narayanan},
url = {http://ict.usc.edu/pubs/SMART%20ROOM-%20PARTICIPANT%20AND%20SPEAKER%20LOCALIZATION%20AND%20IDENTIFICATION.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {Proceedings of the IEEE Conference on Acoustics, Speech and Signal Processing},
address = {Philadelphia, PA},
abstract = {Our long-term objective is to create Smart Room Technologies that are aware of the users presence and their behavior and can become an active, but not an intrusive, part of the interaction. In this work, we present a multimodal approach for estimating and tracking the location and identity of the participants including the active speaker. Our smart room design contains three user-monitoring systems: four CCD cameras, an omnidirectional camera and a 16 channel microphone array. The various sensory modalities are processed both individually and jointly and it is shown that the multimodal approach results in signiï¬cantly improved performance in spatial localization, identiï¬cation and speech activity detection of the participants.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Our long-term objective is to create Smart Room Technologies that are aware of the users presence and their behavior and can become an active, but not an intrusive, part of the interaction. In this work, we present a multimodal approach for estimating and tracking the location and identity of the participants including the active speaker. Our smart room design contains three user-monitoring systems: four CCD cameras, an omnidirectional camera and a 16 channel microphone array. The various sensory modalities are processed both individually and jointly and it is shown that the multimodal approach results in signiï¬cantly improved performance in spatial localization, identiï¬cation and speech activity detection of the participants.
Pair, Jarrell; Treskunov, Anton; Piepol, Diane
Leveraging Hollywood Set Design Techniques to Enhance Ad Hoc Immersive Display Systems Proceedings Article
In: IEEE VR Emerging Displays Workshop, Bonn, Germany, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{pair_leveraging_2005,
title = {Leveraging Hollywood Set Design Techniques to Enhance Ad Hoc Immersive Display Systems},
author = {Jarrell Pair and Anton Treskunov and Diane Piepol},
url = {http://ict.usc.edu/pubs/Leveraging%20Hollywood%20Set%20Design%20Techniques%20to%20Enhance%20Ad%20Hoc%20Immersive%20Display%20Systems.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {IEEE VR Emerging Displays Workshop},
address = {Bonn, Germany},
abstract = {Over the past four years, the FlatWorld project [1] at the University of Southern California Institute for Creative Technologies has exploited ad hoc immersive display techniques to prototype virtual reality education and training applications. While our approach is related to traditional immersive projection systems such as the CAVE [2], our work draws extensively upon techniques widely used in Hollywood sets and theme parks. Our first display system, initially prototyped in 2001, enables wide area virtual environments in which participants can maneuver through simulated rooms, buildings, or streets. In 2004, we expanded our work by experimenting with transparent projection screens. To date, we have used this display technique for presenting life size interactive characters with a pseudo-holographic appearance.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Over the past four years, the FlatWorld project [1] at the University of Southern California Institute for Creative Technologies has exploited ad hoc immersive display techniques to prototype virtual reality education and training applications. While our approach is related to traditional immersive projection systems such as the CAVE [2], our work draws extensively upon techniques widely used in Hollywood sets and theme parks. Our first display system, initially prototyped in 2001, enables wide area virtual environments in which participants can maneuver through simulated rooms, buildings, or streets. In 2004, we expanded our work by experimenting with transparent projection screens. To date, we have used this display technique for presenting life size interactive characters with a pseudo-holographic appearance.
Wang, Dagen; Narayanan, Shrikanth
Speech Rate Estimation Via Temporal Correlation and Selected Sub-band Correlation Proceedings Article
In: Proceedings of the International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pp. 413–416, Philadelphia, PA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{wang_speech_2005,
title = {Speech Rate Estimation Via Temporal Correlation and Selected Sub-band Correlation},
author = {Dagen Wang and Shrikanth Narayanan},
url = {http://ict.usc.edu/pubs/SPEECH%20RATE%20ESTIMATION%20VIA%20TEMPORAL%20CORRELATION%20AND%20SELECTED%20SUB-BAND%20CORRELATION.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {Proceedings of the International Conference on Acoustics, Speech, and Signal Processing (ICASSP)},
pages = {413–416},
address = {Philadelphia, PA},
abstract = {In this paper, we propose a novel method for speech rate estimation without requiring automatic speech recognition. It extends the methods of spectral subband correlation by including temporal correlation and the use of selecting prominent spectral subbands for correlation. Further more, to address some of the practical issues in previously published methods, we introduce some novel components into the algorithm such as the use of pitch confidence, magnifying window, relative peak measure and relative threshold. By selecting the parameters and thresholds from realistic development sets, this method achieves a 0.972 correlation coefficient on syllable number estimation and a 0.706 correlation on speech rate estimation. This result is about 6.9% improvement than current best single estimator and 3.5% improvement than current multi-estimator evaluated on the same switchboard database.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we propose a novel method for speech rate estimation without requiring automatic speech recognition. It extends the methods of spectral subband correlation by including temporal correlation and the use of selecting prominent spectral subbands for correlation. Further more, to address some of the practical issues in previously published methods, we introduce some novel components into the algorithm such as the use of pitch confidence, magnifying window, relative peak measure and relative threshold. By selecting the parameters and thresholds from realistic development sets, this method achieves a 0.972 correlation coefficient on syllable number estimation and a 0.706 correlation on speech rate estimation. This result is about 6.9% improvement than current best single estimator and 3.5% improvement than current multi-estimator evaluated on the same switchboard database.
Peterson, Michael J.; Kyriakakis, Chris
Hybrid Algorithm for Robust, Real-time Source Localization in the Reverberant Environments Proceedings Article
In: International Conference on Acoustics, Speech and Signal Processing, Philadelphia, PA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{peterson_hybrid_2005,
title = {Hybrid Algorithm for Robust, Real-time Source Localization in the Reverberant Environments},
author = {Michael J. Peterson and Chris Kyriakakis},
url = {http://ict.usc.edu/pubs/HYBRID%20ALGORITHM%20FOR%20ROBUST,%20REAL-TIME%20SOURCE%20LOCALIZATION%20IN%20REVERBERANT%20ENVIRONMENTS.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {International Conference on Acoustics, Speech and Signal Processing},
address = {Philadelphia, PA},
abstract = {The location of an acoustical source can be found robustly using the Steered Response Pattern - Phase Transform (SRP-PHAT) algorithm. However SRP-PHAT can be computationally expensive, requiring a search of a large number of candidate locations. The required spacing between these locations is dependent on sampling rate, microphone array geometry, and source location. In this work, a novel method will be presented that calculates a smaller number of test points using an efï¬cient closed-form localization algorithm. This method signiï¬cantly reduces the number of calculations, while still remaining robust in acoustical environments.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The location of an acoustical source can be found robustly using the Steered Response Pattern - Phase Transform (SRP-PHAT) algorithm. However SRP-PHAT can be computationally expensive, requiring a search of a large number of candidate locations. The required spacing between these locations is dependent on sampling rate, microphone array geometry, and source location. In this work, a novel method will be presented that calculates a smaller number of test points using an efï¬cient closed-form localization algorithm. This method signiï¬cantly reduces the number of calculations, while still remaining robust in acoustical environments.
Lee, C. M.; Narayanan, Shrikanth
Toward Detecting Emotions in Spoken Dialogs Proceedings Article
In: IEEE Transactions on Speech and Audio Processing, pp. 293–303, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{lee_toward_2005,
title = {Toward Detecting Emotions in Spoken Dialogs},
author = {C. M. Lee and Shrikanth Narayanan},
url = {http://ict.usc.edu/pubs/Toward%20Detecting%20Emotions%20in%20Spoken%20Dialogs.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {IEEE Transactions on Speech and Audio Processing},
volume = {12},
pages = {293–303},
abstract = {The importance of automatically recognizing emotions from human speech has grown with the increasing role of spoken language interfaces in human-computer interaction applications. This paper explores the detection of domain-specific emotions using language and discourse information in conjunction with acoustic correlates of emotion in speech signals. The specific focus is on a case study of detecting negative and non-negative emotions using spoken language data obtained from a call center application. Most previous studies in emotion recognition have used only the acoustic information contained in speech. In this paper, a combination of three sources of information-acoustic, lexical, and discourse-is used for emotion recognition. To capture emotion information at the language level, an information-theoretic notion of emotional salience is introduced. Optimization of the acoustic correlates of emotion with respect to classification error was accomplished by investigating different feature sets obtained from feature selection, followed by principal component analysis. Experimental results on our call center data show that the best results are obtained when acoustic and language information are combined. Results show that combining all the information, rather than using only acoustic information, improves emotion classification by 40.7% for males and 36.4% for females (linear discriminant classifier used for acoustic information).},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The importance of automatically recognizing emotions from human speech has grown with the increasing role of spoken language interfaces in human-computer interaction applications. This paper explores the detection of domain-specific emotions using language and discourse information in conjunction with acoustic correlates of emotion in speech signals. The specific focus is on a case study of detecting negative and non-negative emotions using spoken language data obtained from a call center application. Most previous studies in emotion recognition have used only the acoustic information contained in speech. In this paper, a combination of three sources of information-acoustic, lexical, and discourse-is used for emotion recognition. To capture emotion information at the language level, an information-theoretic notion of emotional salience is introduced. Optimization of the acoustic correlates of emotion with respect to classification error was accomplished by investigating different feature sets obtained from feature selection, followed by principal component analysis. Experimental results on our call center data show that the best results are obtained when acoustic and language information are combined. Results show that combining all the information, rather than using only acoustic information, improves emotion classification by 40.7% for males and 36.4% for females (linear discriminant classifier used for acoustic information).
Ananthakrishnan, S.; Narayanan, Shrikanth
An Automatic Prosody Recognizer Using a Coupled Multi-stream Acoustic Model and a Syntactic-Prosodic Language Model Proceedings Article
In: Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing, Philadelphia, PA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{ananthakrishnan_automatic_2005,
title = {An Automatic Prosody Recognizer Using a Coupled Multi-stream Acoustic Model and a Syntactic-Prosodic Language Model},
author = {S. Ananthakrishnan and Shrikanth Narayanan},
url = {http://ict.usc.edu/pubs/AN%20AUTOMATIC%20PROSODY%20RECOGNIZER%20USING%20A%20COUPLED%20MULTI-STREAM%20ACOUSTIC%20MODEL%20AND%20A%20SYNTACTIC-PROSODIC%20LANGUAGE%20MODEL.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing},
address = {Philadelphia, PA},
abstract = {Automatic detection and labeling of prosodic events in speech has received much attention from speech technologists and linguists ever since the introduction of annotation standards such as ToBI. Since prosody is intricately bound to the semantics of the utterance, recognition of prosodic events is important for spoken language applications such as automatic understanding and translation of speech. Moreover, corpora labeled with prosodic markers are essential for building speech synthesizers that use data-driven approaches to generate natural speech. In this paper, we build a prosody recognition system that detects stress and prosodic boundaries at the word and syllable level in American English using a coupled Hidden Markov Model (CHMM) to model multiple, asynchronous acoustic feature streams and a syntactic-prosodic model that captures the relationship between the syntax of the utterance and its prosodic structure. Experiments show that the recognizer achieves about 75% agreement on stress labeling and 88% agreement on boundary labeling at the syllable level.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Automatic detection and labeling of prosodic events in speech has received much attention from speech technologists and linguists ever since the introduction of annotation standards such as ToBI. Since prosody is intricately bound to the semantics of the utterance, recognition of prosodic events is important for spoken language applications such as automatic understanding and translation of speech. Moreover, corpora labeled with prosodic markers are essential for building speech synthesizers that use data-driven approaches to generate natural speech. In this paper, we build a prosody recognition system that detects stress and prosodic boundaries at the word and syllable level in American English using a coupled Hidden Markov Model (CHMM) to model multiple, asynchronous acoustic feature streams and a syntactic-prosodic model that captures the relationship between the syntax of the utterance and its prosodic structure. Experiments show that the recognizer achieves about 75% agreement on stress labeling and 88% agreement on boundary labeling at the syllable level.
Hobbs, Jerry R.; Gordon, Andrew S.
Toward a Large-scale Formal Theory of Commonsense Psychology for Metacognition Proceedings Article
In: American Association of Artificial Intelligence Spring Symposium on Metacognitive Computing, Stanford, CA, 2005.
Abstract | Links | BibTeX | Tags: The Narrative Group
@inproceedings{hobbs_toward_2005,
title = {Toward a Large-scale Formal Theory of Commonsense Psychology for Metacognition},
author = {Jerry R. Hobbs and Andrew S. Gordon},
url = {http://ict.usc.edu/pubs/Toward%20a%20Large-scale%20Formal%20Theory%20of%20Commonsense%20Psychology%20for%20Metacognition.pdf},
year = {2005},
date = {2005-03-01},
booktitle = {American Association of Artificial Intelligence Spring Symposium on Metacognitive Computing},
address = {Stanford, CA},
abstract = {Robust intelligent systems will require a capacity for metacognitive reasoning, where intelligent systems monitor and reflect on their own reasoning processes. A large-scale study of human strategic reasoning indicates that rich representational models of commonsense psychology are available to enable human metacognition. In this paper, we argue that large-scale formalizations of commonsense psychology enable metacognitive reasoning in intelligent systems. We describe our progress toward developing 30 integrated axiomatic theories of commonsense psychology, and discuss the central representational challenges that have arisen in this work to date.},
keywords = {The Narrative Group},
pubstate = {published},
tppubtype = {inproceedings}
}
Robust intelligent systems will require a capacity for metacognitive reasoning, where intelligent systems monitor and reflect on their own reasoning processes. A large-scale study of human strategic reasoning indicates that rich representational models of commonsense psychology are available to enable human metacognition. In this paper, we argue that large-scale formalizations of commonsense psychology enable metacognitive reasoning in intelligent systems. We describe our progress toward developing 30 integrated axiomatic theories of commonsense psychology, and discuss the central representational challenges that have arisen in this work to date.
Grote, Christopher L.; Parsons, Thomas D.
Threats to the Livelihood of the Forensic Neuropsychological Practice: Avoiding Ethical Misconduct Journal Article
In: Journal of Forensic Neuropsychology, vol. 4, no. 3, pp. 79–93, 2005.
Abstract | Links | BibTeX | Tags: MedVR
@article{grote_threats_2005,
title = {Threats to the Livelihood of the Forensic Neuropsychological Practice: Avoiding Ethical Misconduct},
author = {Christopher L. Grote and Thomas D. Parsons},
url = {http://ict.usc.edu/pubs/Threats%20to%20the%20Livelihood%20of%20the%20Forensic%20Neuropsychological%20Practice-%20Avoiding%20Ethical%20Misconduct.pdf},
doi = {10.1300/J151v04n03_06},
year = {2005},
date = {2005-01-01},
journal = {Journal of Forensic Neuropsychology},
volume = {4},
number = {3},
pages = {79–93},
abstract = {We review six diverse issues that have the potential of devaluing our profession, in that ethical missteps could lead to the perception or reality that the work of forensic neuropsychologists is "for sale." By resisting temptations or overtures to engage in inappropriate conduct, such as attacking colleagues or failing to recognize how our own biases might influence our behavior or opinions, neuropsychologists strive to create a work product that enhances the reputation of our profession and makes a positive contribution to the public-at-large.},
keywords = {MedVR},
pubstate = {published},
tppubtype = {article}
}
We review six diverse issues that have the potential of devaluing our profession, in that ethical missteps could lead to the perception or reality that the work of forensic neuropsychologists is "for sale." By resisting temptations or overtures to engage in inappropriate conduct, such as attacking colleagues or failing to recognize how our own biases might influence our behavior or opinions, neuropsychologists strive to create a work product that enhances the reputation of our profession and makes a positive contribution to the public-at-large.
Martinovski, Bilyana; Traum, David; Marsella, Stacy C.
Rejection of empathy and its linguistic manifestations Proceedings Article
In: Proceedings of Conference on Formal and Informal Negotiation (FINEXIN), Ottowa, Canada, 2005.
Abstract | Links | BibTeX | Tags: Social Simulation, Virtual Humans
@inproceedings{martinovski_rejection_2005,
title = {Rejection of empathy and its linguistic manifestations},
author = {Bilyana Martinovski and David Traum and Stacy C. Marsella},
url = {http://ict.usc.edu/pubs/Rejection%20of%20empathy%20and%20its%20linguistic%20manifestations.pdf},
year = {2005},
date = {2005-01-01},
booktitle = {Proceedings of Conference on Formal and Informal Negotiation (FINEXIN)},
address = {Ottowa, Canada},
abstract = {Trust is a crucial quality in the development of individuals and societies and empathy plays a key role in the formation of trust. Trust and empathy have growing importance in studies of negotiation. However, empathy can be rejected which complicates its role in negotiation. This paper presents a linguistic analysis of empathy by focusing on rejection of empathy in negotiation. Some of the rejections are due to failed recognition of the rejector's needs and desires whereas others have mainly strategic functions gaining momentum in the negotiation. In both cases, rejection of empathy is a phase in the negotiation not a breakdown.},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {inproceedings}
}
Trust is a crucial quality in the development of individuals and societies and empathy plays a key role in the formation of trust. Trust and empathy have growing importance in studies of negotiation. However, empathy can be rejected which complicates its role in negotiation. This paper presents a linguistic analysis of empathy by focusing on rejection of empathy in negotiation. Some of the rejections are due to failed recognition of the rejector's needs and desires whereas others have mainly strategic functions gaining momentum in the negotiation. In both cases, rejection of empathy is a phase in the negotiation not a breakdown.
Parsons, Thomas D.; Rizzo, Albert; Zaag, Cheryl; McGee, Jocelyn; Buckwalter, John Galen
Gender Differences and Cognition Among Older Adults Journal Article
In: Aging, Neuropsychology, and Cognition, vol. 12, pp. 78–88, 2005.
Abstract | Links | BibTeX | Tags: MedVR
@article{parsons_gender_2005,
title = {Gender Differences and Cognition Among Older Adults},
author = {Thomas D. Parsons and Albert Rizzo and Cheryl Zaag and Jocelyn McGee and John Galen Buckwalter},
url = {http://ict.usc.edu/pubs/Gender%20Differences%20and%20Cognition%20Among%20Older%20Adults.pdf},
year = {2005},
date = {2005-01-01},
journal = {Aging, Neuropsychology, and Cognition},
volume = {12},
pages = {78–88},
abstract = {The more replicated findings about gender difference in cognitive performance suggest female superiority on visuomotor speed and language ability and male superiority on mechanical and visuospatial tasks. Generally, group strengths found in the early school years become more established at adolescence and remain stable through adulthood. The current study tested whether the patterns established in the early years remained among 30 adult subjects. We also utilized a series of exploratory analyses to determine if observed gender differences were impacted by the covariance present between all cognitive tests. Results suggest that although the patterns established in the early years remain stable through time for males, the established patterns for females are altered with age. Our findings are compelling in supporting a male advantage on visuospatial tasks among older adults. These findings are discussed in terms of common variance between test instruments as a possible source of difference. Our finding that the gender effect tended to increase when common variance was controlled argues that this methodology may enhance the ability to detect domain specific effects.},
keywords = {MedVR},
pubstate = {published},
tppubtype = {article}
}
The more replicated findings about gender difference in cognitive performance suggest female superiority on visuomotor speed and language ability and male superiority on mechanical and visuospatial tasks. Generally, group strengths found in the early school years become more established at adolescence and remain stable through adulthood. The current study tested whether the patterns established in the early years remained among 30 adult subjects. We also utilized a series of exploratory analyses to determine if observed gender differences were impacted by the covariance present between all cognitive tests. Results suggest that although the patterns established in the early years remain stable through time for males, the established patterns for females are altered with age. Our findings are compelling in supporting a male advantage on visuospatial tasks among older adults. These findings are discussed in terms of common variance between test instruments as a possible source of difference. Our finding that the gender effect tended to increase when common variance was controlled argues that this methodology may enhance the ability to detect domain specific effects.
Robertson, R. Kevin; Mielke, Jens; Appiah, Kuku; Hall, Colin D.; Price, Richard W.; Kumwenda, Johnstone; Kanyama, Cecelia; Amod, Farida; Marra, Christina; Taylor, Terrie; Lalloo, Umesh; Jelsma, Jennifer; Holding, Penny; Boivin, Michael; Birbeck, Gretchen; Nakasujja, Noeline; Sanne, Ian; Parsons, Thomas D.; Parente, Amanda; Tucker, Karen A.
Assessment of neuroAIDS in Africa Journal Article
In: Journal of NeuroVirology, vol. 11, no. S1, pp. 7–16, 2005.
Abstract | Links | BibTeX | Tags: MedVR
@article{robertson_assessment_2005,
title = {Assessment of neuroAIDS in Africa},
author = {R. Kevin Robertson and Jens Mielke and Kuku Appiah and Colin D. Hall and Richard W. Price and Johnstone Kumwenda and Cecelia Kanyama and Farida Amod and Christina Marra and Terrie Taylor and Umesh Lalloo and Jennifer Jelsma and Penny Holding and Michael Boivin and Gretchen Birbeck and Noeline Nakasujja and Ian Sanne and Thomas D. Parsons and Amanda Parente and Karen A. Tucker},
url = {http://ict.usc.edu/pubs/Assessment%20of%20neuroAIDS%20in%20Africa.pdf},
year = {2005},
date = {2005-01-01},
journal = {Journal of NeuroVirology},
volume = {11},
number = {S1},
pages = {7–16},
abstract = {In June of 2004, the Center for AIDS Mental Health Research, National Institute of Mental Health sponsored a conference on the Assessment of NeuroAIDS in Africa, which was held in Blantrye, Malawai. The conference presentations summarized here highlight the need for research on NeuroAIDS in Africa and methods for assessing HIV-related neurological diseases (Robertson, 2004).},
keywords = {MedVR},
pubstate = {published},
tppubtype = {article}
}
In June of 2004, the Center for AIDS Mental Health Research, National Institute of Mental Health sponsored a conference on the Assessment of NeuroAIDS in Africa, which was held in Blantrye, Malawai. The conference presentations summarized here highlight the need for research on NeuroAIDS in Africa and methods for assessing HIV-related neurological diseases (Robertson, 2004).
Gratch, Jonathan; Marsella, Stacy C.
Lessons from Emotion Psychology for the Design of Lifelike Characters Journal Article
In: Applied Artificial Intelligence Journal, vol. 19, pp. 215–233, 2005.
Abstract | Links | BibTeX | Tags: Social Simulation, Virtual Humans
@article{gratch_lessons_2005,
title = {Lessons from Emotion Psychology for the Design of Lifelike Characters},
author = {Jonathan Gratch and Stacy C. Marsella},
url = {http://ict.usc.edu/pubs/Lessons%20from%20Emotion%20Psychology%20for%20the%20Design%20of%20Lifelike%20Characters.pdf},
year = {2005},
date = {2005-01-01},
journal = {Applied Artificial Intelligence Journal},
volume = {19},
pages = {215–233},
abstract = {This special issue describes a number of applications that utilize lifelike characters that teach indirectly, by playing some role in a social interaction with a user. The design of such systems reflects a compromise between competing, sometimes unarticulated de- mands: they must realistically exhibit the behaviors and characteristics of their role, they must facilitate the desired learning, and they must work within the limitations of current technology, and there is little theoretical or empirical guidance on the impact of these compromises on learning. Our perspective on this problem is shaped by our interest in the role of emotion and emotional behaviors in such forms of learning. In recent years, there has been an explosion of interest in the role of emotion in the design of virtual hu- mans. The techniques and motivations underlying these various efforts can seem, from an outsider's perspective, as bewildering and multifaceted as the concept of emotion itself is generally accused of being. Drawing on insights from emotion psychology, this article attempts to clarify for the designers of educational agents the various theoretical perspec- tives on the concept of emotion with the aim of giving guidance to designers of educa- tional agents.},
keywords = {Social Simulation, Virtual Humans},
pubstate = {published},
tppubtype = {article}
}
This special issue describes a number of applications that utilize lifelike characters that teach indirectly, by playing some role in a social interaction with a user. The design of such systems reflects a compromise between competing, sometimes unarticulated de- mands: they must realistically exhibit the behaviors and characteristics of their role, they must facilitate the desired learning, and they must work within the limitations of current technology, and there is little theoretical or empirical guidance on the impact of these compromises on learning. Our perspective on this problem is shaped by our interest in the role of emotion and emotional behaviors in such forms of learning. In recent years, there has been an explosion of interest in the role of emotion in the design of virtual hu- mans. The techniques and motivations underlying these various efforts can seem, from an outsider's perspective, as bewildering and multifaceted as the concept of emotion itself is generally accused of being. Drawing on insights from emotion psychology, this article attempts to clarify for the designers of educational agents the various theoretical perspec- tives on the concept of emotion with the aim of giving guidance to designers of educa- tional agents.
Hawkins, Tim; Einarsson, Per; Debevec, Paul
A Dual Light Stage Proceedings Article
In: Dutré, Philip; Bala, Kavita (Ed.): Eurographics Symposium on Rendering, Konstanz, Germany, 2005.
Abstract | Links | BibTeX | Tags: Graphics
@inproceedings{hawkins_dual_2005,
title = {A Dual Light Stage},
author = {Tim Hawkins and Per Einarsson and Paul Debevec},
editor = {Philip Dutré and Kavita Bala},
url = {http://ict.usc.edu/pubs/A%20Dual%20Light%20Stage.pdf},
year = {2005},
date = {2005-01-01},
booktitle = {Eurographics Symposium on Rendering},
address = {Konstanz, Germany},
abstract = {We present a technique for capturing high-resolution 4D reflectance ï¬elds using the reciprocity property of light transport. In our technique we place the object inside a diffuse spherical shell and scan a laser across its surface. For each incident ray, the object scatters a pattern of light onto the inner surface of the sphere, and we photograph the resulting radiance from the sphere's interior using a camera with a ï¬sheye lens. Because of reciprocity, the image of the inside of the sphere corresponds to the reflectance function of the surface point illuminated by the laser, that is, the color that point would appear to a camera along the laser ray when the object is lit from each direction on the surface of the sphere. The measured reflectance functions allow the object to be photorealistically rendered from the laser's viewpoint under arbitrary directional illumination conditions. Since each captured reflectance function is a high-resolution image, our data reproduces sharp specular reflections and self-shadowing more accurately than previous approaches. We demonstrate our technique by scanning objects with a wide range of reflectance properties and show accurate renderings of the objects under novel illumination conditions.},
keywords = {Graphics},
pubstate = {published},
tppubtype = {inproceedings}
}
We present a technique for capturing high-resolution 4D reflectance ï¬elds using the reciprocity property of light transport. In our technique we place the object inside a diffuse spherical shell and scan a laser across its surface. For each incident ray, the object scatters a pattern of light onto the inner surface of the sphere, and we photograph the resulting radiance from the sphere's interior using a camera with a ï¬sheye lens. Because of reciprocity, the image of the inside of the sphere corresponds to the reflectance function of the surface point illuminated by the laser, that is, the color that point would appear to a camera along the laser ray when the object is lit from each direction on the surface of the sphere. The measured reflectance functions allow the object to be photorealistically rendered from the laser's viewpoint under arbitrary directional illumination conditions. Since each captured reflectance function is a high-resolution image, our data reproduces sharp specular reflections and self-shadowing more accurately than previous approaches. We demonstrate our technique by scanning objects with a wide range of reflectance properties and show accurate renderings of the objects under novel illumination conditions.
Kallmann, Marcelo
Scalable Solutions for Interactive Virtual Humans that can Manipulate Objects Proceedings Article
In: First Annual Artificial Intelligence and Interactive Entertainment Conference, Marina del Rey, CA, 2005.
Abstract | Links | BibTeX | Tags:
@inproceedings{kallmann_scalable_2005,
title = {Scalable Solutions for Interactive Virtual Humans that can Manipulate Objects},
author = {Marcelo Kallmann},
url = {http://ict.usc.edu/pubs/Scalable%20Solutions%20for%20Interactive%20Virtual%20Humans%20that%20can%20Manipulate%20Objects.pdf},
year = {2005},
date = {2005-01-01},
booktitle = {First Annual Artificial Intelligence and Interactive Entertainment Conference},
address = {Marina del Rey, CA},
abstract = {This paper presents scalable solutions for achieving virtual humans able to manipulate objects in interactive virtual environments. The scalability trades computational time with the ability of addressing increasingly difficult constraints. In time-critical environments, arm motions are computed in few milliseconds using fast analytical Inverse Kinematics. For other types of applications where collision-free motions are required, a randomized motion planner capable of generating motions of average complexity in about a second of computation time is employed. The steps required for defining and computing different types of manipulations are described in this paper.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper presents scalable solutions for achieving virtual humans able to manipulate objects in interactive virtual environments. The scalability trades computational time with the ability of addressing increasingly difficult constraints. In time-critical environments, arm motions are computed in few milliseconds using fast analytical Inverse Kinematics. For other types of applications where collision-free motions are required, a randomized motion planner capable of generating motions of average complexity in about a second of computation time is employed. The steps required for defining and computing different types of manipulations are described in this paper.