Publications

@inproceedings{chen_dont_nodate,

title = {Don't Look into the Dark: Latent Codes for Pluralistic Image Inpainting},

author = {Haiwei Chen and Yajie Zhao},

url = {https://openaccess.thecvf.com/content/CVPR2024/html/Chen_Dont_Look_into_the_Dark_Latent_Codes_for_Pluralistic_Image_CVPR_2024_paper.html},

pages = {7591–7600},

abstract = {We present a method for large-mask pluralistic image inpainting based on the generative framework of discrete latent codes. Our method learns latent priors discretized as tokens by only performing computations at the visible locations of the image. This is realized by a restrictive partial encoder that predicts the token label for each visible block a bidirectional transformer that infers the missing labels by only looking at these tokens and a dedicated synthesis network that couples the tokens with the partial image priors to generate coherent and pluralistic complete image even under extreme mask settings. Experiments on public benchmarks validate our design choices as the proposed method outperforms strong baselines in both visual quality and diversity metrics.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{artstein_augmenting_nodate,

title = {Augmenting Training Data for a Virtual Character Using GPT-3.5},

author = {Ron Artstein and Elizabeth Chen},

url = {https://journals.flvc.org/FLAIRS/article/view/135552},

volume = {37},

publisher = {Tyhe Florida Artificial Intelligence Research Society},

abstract = {This paper compares different methods of using a large lan-guage model (GPT-3.5) for creating synthetic training datafor a retrieval-based conversational character. The trainingdata are in the form of linked questions and answers, whichallow a classifier to retrieve a pre-recorded answer to an un-seen question; the intuition is that a large language modelcould predict what human users might ask, thus saving theeffort of collecting real user questions as training data. Re-sults show small improvements in test performance for allsynthetic datasets. However, a classifier trained on only smallamounts of collected user data resulted in a higher F-scorethan the classifiers trained on much larger amounts of syn-thetic data generated using GPT-3.5. Based on these results,we see a potential in using large language models for gener-ating training data, but at this point it is not as valuable ascollecting actual user data for training.},

keywords = {},

pubstate = {published},

tppubtype = {inproceedings}

}