Research Experience for Undergraduates (REU)

Applications for summer 2027 are open from November 9, 2026 to January 27, 2027

The Institute for Creative Technologies (ICT) offers a 10-week summer research program for undergraduates in intelligent interactive experiences. A multidisciplinary research institute affiliated with the University of Southern California, the ICT was established in 1999 to combine leading academic researchers in computing with the creative talents of Hollywood and the video game industry. Having grown to encompass over 130 faculty, staff, and students in a diverse array of fields, the ICT represents a unique interdisciplinary community brought together with a core unifying mission: advancing the state-of-the-art for creating virtual experiences so compelling that people will react as if they were real.

Reflecting the interdisciplinary nature of ICT research, we welcome applications from students in computer science, as well as many other fields, such as psychology, art/animation, interactive media, linguistics, and communications. Undergraduates will join a team of students, research staff, and faculty in one of several labs focusing on different aspects of interactive virtual experiences. In addition to participating in seminars and social events, students will also prepare a final written report and present their projects to the rest of the institute at the end of the summer.

This Research Experiences for Undergraduates (REU) site is supported by a grant from the National Science Foundation.

ICT also offers another internship program, for both undergraduate and graduate students, which requires a separate application, view a list of available positions here. For questions or additional information, please contact [email protected].

Natural Language Dialogue Processing

Mentors: David Traum, Ron Artstein and Kallirroi Georgila

The Natural Language Dialogue Group at ICT is developing artificial intelligence and language technology to allow machines to participate in human-like natural dialogues with people. Our systems include virtual humans, robots, recorded real people, and voice and chat systems. An REU student will work on creating, extending or evaluating such systems; analyzing conversational data or collecting new data; or other topics using dialogue data and state-of-the-art machine learning methods. Specific projects can be chosen or defined by the REU student; examples include technology to allow an agent or robot to understand the context of a conversation, take initiative, or sustain interaction over multiple encounters; use of large language models for system response generation; and reinforcement learning of system policies. Previous REU students in the group have been lead authors and had their REU projects published in the proceedings of international scientific conferences.

AI Agents for Analyzing and Mitigating Cross-cultural Conflict

Mentors: Gale Lucas

Conflicts evoke emotion and often arise from differences in underlying personal and cultural values. Our research uses AI to gain insight into how culture shapes disputes, and develops methods to teach and coach conflict-resolution skills. We adopt a multidisciplinary approach that combines theories from psychology and dispute resolution with artificial intelligence methods (e.g., use of machine learning and large language models). Students will work with a large multicultural corpus of video disputes. REU students are sought to assist with developing research software, experiments and data analysis. Candidates should have good programming skills. Prior experience with large-language models, including fine-tuning and reinforcement learning is a benefit. Fluency in Mandarin is also a benefit.

Artificial Intelligence for Education

Mentor: Benjamin Nye

This project will explore applications of Artificial Intelligence and Machine Learning to support personalized tutoring and dialog systems such as the AI for Brainstorming and Editing (ABE), Personal Assistant for Life-Long Learning (PAL3) or CareerFair.ai virtual mentors (MentorPal). These systems use AI agents and frameworks for tasks such as analyzing and improving student writing, delivering engaging and accessible educational content, and offering career mentoring. Interested students are encouraged to explore how to use multi-agent and agentic AI frameworks, such as AG2.

The goal of the REU internship will be to expand the repertoire of the system to further enhance learning and engagement. The specific tasks will be determined collaboratively based on research goals and student research interests. Students will be expected to contribute to a peer-reviewed publication.

Possible topics include
(1)  Working with Generative AI models for developing intelligent tutoring content;
(2) Leveraging multimodal models in learning technologies,
(3) Modifying the intelligent tutoring system and how it supports the learner, and
(4) Statistical analysis, and/or data mining to identify patterns of interactions between human subjects and the intelligent tutoring system.

Benchmarking Physical Consistency in Generative Video and World Models with 3D Ground Truth Data

Mentor: Yajie Zhao

Generative video and world models produce visually convincing scenes, but it is unclear whether they have learned how the real world truly behaves. Visual realism alone cannot answer this question, and common perceptual metrics may reward results that look convincing but are physically wrong. Do they get interpersonal interactions, behaviors, shadows, and interreflections right when illumination changes, or do they only reproduce plausible-looking videos? Answering this requires ground truth that is scarcely available: the same subjects and objects captured in many different views under precisely known, controlled lighting. In this project, the intern will use our light stage to build a small benchmark of objects and subjects captured in different states of action, behavior and illumination. Because light transport is linear, these captures let us synthesize physically exact images under arbitrary lighting, including novel environment maps and moving lights. This gives us a reference for what the correct response to a lighting change looks like. The intern will then evaluate existing image and video generation models by prompting or conditioning them with a starting frame and a lighting change. The outputs are compared against ground truth using metrics the intern will help design. These may include image-based error measures, shadow and highlight consistency, temporal stability under moving lights, and preservation of subject identity, geometry, and motion. To help separate model limitations from ambiguity in the conditioning, tests will range from simple changes, such as moving a single light, to complex changes between environments. At the end of the project, we will have an open evaluation suite consisting of the benchmark dataset, ground-truth synthesis pipeline and metrics, along with a ranking of current models by physical lighting fidelity.

Machine Learning for Human Behavior Understanding

Mentor: Mohammad Soleymani

The Intelligent Human Perception Lab at USC’s Institute for Creative Technologies conducts research on automatic human behavior analysis. To model human emotions and behaviors, we research machine learning methods that can fuse multiple modalities and generalize to new people. We are seeking REU interns to work on related research tasks, including the following topics: multimodal LLM for behavior understanding, multimodal learning for emotion recognition and sentiment analysis and generative AI for human behavior generation. REU interns who are interested in these topics should be comfortable handling data and able to program in Python. Prior experience with machine learning frameworks, e.g., PyTorch, is a plus.

Human-inspired Adaptive Teaming Systems

Mentor: Volkan Ustun

AI agents are increasingly expected to work in teams with other agents and with people, from military training and emergency-response simulations to multiplayer games and robotics. Human teams often coordinate with limited communication, and building AI agents that can do the same is a challenge. We approach it by combining multi-agent reinforcement learning with generative models, such as large language models and diffusion models, using simulated environments and data from multiplayer games. Student researchers will build or extend simulation environments, design scenarios, collect and analyze multi-agent data, and train and evaluate models.

Paradigms for Strategic Human-machine Integration

Mentor: Nikolos Gurney

Through a hypothesis-driven methodology, we will investigate the cognitive, computational, and social mechanisms unique to HMI, develop novel measurement approaches specifically designed for human-machine dynamics, and construct integrative models that optimize these partnerships. Our central hypothesis is that purpose-built HMI frameworks will demonstrate superior predictive power across diverse operational scenarios compared to adapted human team models.