The Vision, Language, and Learning Lab at the University of Virginia

News and Announcements

03/2023. CVPR 2024: Moayed has his work on arbitrary size text-to-image generation accepted [ElasticDiffusion] and Catherine has her work on visual grounding with self consistent predictions for equivalent phrases accepted [SelfEQ].
12/2023. Ziyan has her work on visual relationship grounding accepted to WACV 2024 [SCoRD] and successfully defended her PhD Defense!.
07/2023. Paola has her work on vision-and-language beyond nouns accepted to ICCV 2023, Ziyan has her work on visual grounding accepted to CVPR 2023, and Aman has CLIP-Lite accepted to AISTATS 2023.
08/2022. We receive a Google Inclusion Research Award 2022.
04/2022. Paola and Letao have SimVQA accepted to CVPR 2022, and Ziyan has Backpropagation-based decoding for MMT accepted to Frontiers in AI.
07/2021. Two papers accepted to ICCV 2021, Reranking Transformers [arxiv] and MEDIRL [arxiv].
07/2021. After some wonderful five years at the University of Virginia, our group is in the process of moving to the Department of Computer Science at Rice University in Houston, Texas~!
06/2021. Our work on teaching machines compositional vision and language models is funded through a National Science Foundation CAREER Award
06/2021. Tianlu Wang defends her PhD Dissertation Measuring and Mitigating Biases in Vision and Language Models, accepts position as Research Scientist at Facebook AI Research in Menlo Park, California. Congrats Tianlu~!
04/2021. Fuwen Tan defends his PhD Dissertation Learning Local Representations of Images and Text, accepts position at Samsung AI Center - Cambridge. Congrats Fuwen~!
03/2021. Two papers accepted to CVPR 2021, Black-box Explanation of Object Detectors, and the Classification Transformer.
02/2021. Our group received a Salesforce AI Research Grant and an NSF-Amazon Fairness in AI Grant to support our work.
12/2020. Paola and Fuwen had Curriculum Labeling accepted to AAAI.
09/2020. Tianlu, Ziyan and Leticia got papers accepted to EMNLP and Findings of EMNLP 2020 [publications].
06/2020. Andrew Ng's deeplearning.ai has a blog post highlighting our ACL 2020 paper with Salesforce Research on Double-Hard Debias [link].
06/2020. TechXplore features our work on regulation of Face Recognition in Researchers call for new federal authority to regulate facial recognition tech.
05/2020. With a group of colleagues and funding from the MacArthur Foundation, we released the whitepaper Face Recognition Technologies in the Wild: A Call for a Federal Office.
05/2020. Our group recently received a Facebook Research Award 2020 and gift funding from eBay Research and Adobe Research.
04/2020. Our vislang.ai website is up and running entirely on the cloud!
04/2020. New papers accepted at CVPR 2020, ACL 2020, and ICSE 2020.
02/2020. Co-Organizing with colleagues at the Seoul National University and other places, the 2nd workshop on Video Turing Test: Toward Human-Level Video Story Understanding at ECCV 2020. Send your submissions to the DramaQA Challenge and attend our workshop!
07/2019. Posts from NVIDIA [link] and IBM Research [link] about Text2Scene
07/2019. Text2Scene gets named among 45 Best CVPR Paper Finalists among 1,294 accepted papers (top 1% of all submissions) [link]
05/2019. UVA Today features work by PhD student Tianlu Wang under her Presidential Fellowship Using Data Science to determine why one job candidate beats out another.
05/2019. Participated at the Ethics in AI Panel at Escape Velocity 2019 that took place in Washington DC's National Harbor.
04/2019. PhD Student Tianlu Wang gave a talk at the TomTom Applied Machine Learning Conference.
09/2018. Co-organizing and Participating in the panel on Dealing with Bias and Unfairness in ML at the ACM Richard Tapia Celebration of Diversity in Computing, Orlando, FL.
02/2018. Received a Google Faculty Research Award and an IBM Faculty Award.
08/2017. Our work at UVA with UCLA's NLP Group gets coverage in WIRED, Daily Mail, The Times of London, Glamour, Bloomberg, among others.
09/2017. We obtained a Best --Long-- Paper Award at EMNLP 2017~!

This demo attemps to highlight areas of an image conditioned on an arbitrary input text.

This demo attemps to make it difficult for a model to predict gender from an image by modifying it so that this task becomes harder while retaining most image information.

Text2Scene

This demo turns textual descriptions into a scene generated automatically by stitching objects sequentially on a plain background step-by-step using sequence generation neural networks.

Visual Translator

This demo attemps to translate a sentence in English into visual feature space and into a sentence in both German (Deutsch) and Japanese (日本語).

CURRENT AND PAST SPONSORS