This is the first workshop on domain adaptation and robust learning across geographies in computer vision. This workshop aims to unite researchers from across the vision community to foster discussions on the spectrum of challenges posed by geographic bias towards fair and inclusive computer vision. Our primary objective is to facilitate discussions and ideas that enable effective deployment of modern AI technology in low and mid-income societies through the design of geographically robust and transferable models that counter such biases. As part of this workshop, we are also conducting a challenge to benchmark progress in learning geographically robust classification models designed to highlight the emerging open problems in this area. Tapping on the recently introduced large-scale geographic adaptation dataset GeoNet, we host three challenges on unsupervised domain adaptation, single source domain generalization and universal domain adaptation. The workshop consists of presentations by experts in the field and short talks regarding methods that successfully address the challenges.
1st GeoNet Challenge on Geographical Robustness
The 1st GeoNet challenge contains three tracks, aimed at tackling different problems in geographical robustness. All the tracks will use the recently proposed GeoNet dataset for training and validation. Test set will be separately released later. More details about the challenge, inlcuding general rules and guidelines, are presented on the challenge page.
- Unsupervised Domain Adaptation: is targeted to improve accuracy on an unlabeled target geography using labeled images from a different source geography.
- Single Source Domain Generalization:, where the task is to leverage images from a specific source geography and generalize well to unseen test geographies.
- Universal Domain Adaptation:, where the aim is to generalize to the target domain with unlabeled samples during training and novel categories at test-time, while using labels from a separate source geography.
|Challenge Data Released||June 15, 2023|
|Challenge Test Set Released||August 1, 2023|
|Submission Portal Open||Aug 10, 2023|
|Challenge Submission Deadline||Aug 20, 2023|
|Workshop Date||Oct 2, 2023 (Day 1 of ICCV 2023)|
Dr. Kristen Grauman is a Professor in the Department of Computer Science at the University of Texas at Austin. Her primary research interests are visual recognition and visual search and she is well-known for her pioneering works on large-scale image/video retrieval, active learning, active recognition, first-person "egocentric" computer vision, multimodal learning, activity recognition, vision and language, and video summarization. She has also recently led a team of researchers worldwide towards developing a planet-scale egocentric video dataset called Ego4D, which had a significant influence on the landscape of egocentric computer vision.
Dr. Judy Hoffman is an Assistant Professor in the School of Interactive Computing at Georgia Tech, a member of the Machine Learning Center, and a Diversity and Inclusion Fellow. Her research lies at the intersection of computer vision and machine learning with specialization in domain adaptation, transfer learning, adversarial robustness, and algorithmic fairness. She has received numerous awards including NSF CAREER, Google Research Scholar Award (2022), Samsung AI Researcher of the Year Award (2021), NVIDIA female leader in computer vision award (2020), AIMiner top 100 most influential scholars in Machine Learning (2020), MIT EECS Rising Star in 2015, and the NSF Graduate Fellowship.
Dr. Carl Vondrick is an associate professor of computer science at Columbia University. His research focuses on computer vision and machine learning. His research is supported by the NSF, DARPA, Amazon, and Toyota, and his work has appeared on the national news, such as CNN, NPR, the Associated Press and Stephen Colbert's television show. He received the 2021 NSF CAREER Award, the 2021 Toyota Young Faculty Award, and the 2018 Amazon Research Award. Previously, he was a Research Scientist at Google and he received his PhD from MIT in 2017.
Dr. Sara Beery is currently a Visiting Researcher at Google working on urban forest monitoring, and will join MIT as an assistant professor in the Faculty of Artificial Intelligence and Decision-Making in EECS in September 2023. Beery received her PhD in computing and mathematical sciences at Caltech in 2022, where she was advised by Pietro Perona. Her research focuses on building computer vision methods that enable global-scale environmental and biodiversity monitoring across data modalities, tackling real-world challenges including strong spatiotemporal correlations, imperfect data quality, fine-grained categories, and long-tailed distributions. She partners with nongovernmental organizations and government agencies to deploy her methods in the wild worldwide and works toward increasing the diversity and accessibility of academic research in artificial intelligence through interdisciplinary capacity building and education.
Dr. Girmaw Abebe Tadesse is a Principal Research Scientist and Manager at Microsoft AI for Good Research Lab which aims to develop AI solutions for critical problems across sectors including agriculture, healthcare, biodiversity, etc. Previously at IBM Research Africa, Girmaw led multiple projects in trustworthy AI including evaluation of generative models, representation analysis in academic materials and data-driven insight extraction from public healthy surveys, with active collaborations with external institutions such as Bill & Melinda Gates Foundation, Stanford University, Oxford University and Harvard University. He has interned/worked in various research groups across Europe, including the UPC-BarcelonaTech (Spain), KU Leuven (Belgium), and INESC-ID (Portugal).