Dear all,
Learning from human preferences has been critical to major recent advances in AI and machine learning, such as fine-tuning of large language models, robotics, and self-driving cars. While the recent advances were very visual and captured our imagination, the core ideas behind them have been percolating through our communities for a long time, including in - Collaborative filtering - Control theory - Convex optimization - Dueling and preference-based bandits - Econometrics and assortment selection - Explainability - Fairness - Game theory, equilibria, and multiplayer games - Learning from human and AI feedback - Marketing and revenue management - Multi-objective optimization - Preference elicitation - Ranking aggregation - Recommender systems - Reinforcement learning - Robotics - Search engine optimization - Social choice theory To bring all ideas together, and shape the future of this area, we wanted to invite you to submit your work to ICML 2023 Workshop “The Many Facets of Preference-Based Learning”. The workshop will be held in Hawaii on July 28, 2023. The paper submission deadline is in less than 3 weeks and you can find more details below. Sincerely, Viktor Bengs (LMU, Germany) Robert Busa-Fekete (Google Research) Mohammad Ghavamzadeh (Google Research) Branislav Kveton (AWS AI Labs) Aadirupa Saha (Apple Research) - - - - - - - - - - CALL FOR PAPERS ICML 2023 Workshop: The Many Facets of Preference-Based Learning Honolulu, Hawaii HOMEPAGE: https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsites.google.com%2Fview%2Fmfpl-icml-2023&data=05%7C01%7Cuai%40engr.orst.edu%7C658116502d6c49af055f08db6b5d5415%7Cce6d05e13c5e4d6287a84c4a2713c113%7C0%7C0%7C638221820814333160%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=2Ol%2FhZzR1DwBKkPXvFpX8Kpxl0upokddLMszF%2FItjU0%3D&reserved=0 EMAIL: learningprefere...@gmail.com IMPORTANT DATES Paper Submission Deadline: June 16, 2023 Notification of Acceptance: June 26, 2023 Workshop: July 28, 2023 WORKSHOP OVERVIEW Learning from human feedback has become increasingly important as the complexity of problems solved by AI and machine learning grows. While humans often find it difficult to provide demonstrations of the desired system’s behavior or to quantify its responses using numerical values, providing preferences (or comparisons) is natural. Therefore, it is not surprising that learning from human preferences has been critical to major recent advances in AI and machine learning, such as fine-tuning of large language models, guided image generation, robotics, and self-driving cars. Despite these ground-breaking successes, the most exciting opportunities still lie ahead of us. The goal of this workshop is to bring together scientists from communities where preference-based learning has played a major role or has a potential for making a breakthrough. We want to celebrate recent advances, discuss main challenges and potential solutions, and pave the way for future research directions. Additionally, we aim to strengthen the connection between theory and practice by identifying real-world systems that can benefit from incorporating preference feedback. We cordially invite scientists who feel addressed by the theme of the workshop to submit their latest works. Since preference-based learning had impact on many communities, potential topics could be, but are not limited to, - Collaborative filtering - Control theory - Convex optimization - Dueling and preference-based bandits - Econometrics and assortment selection - Explainability - Fairness - Game theory, equilibria, and multiplayer games - Learning from human and AI feedback - Marketing and revenue management - Multi-objective optimization - Preference elicitation - Ranking aggregation - Recommender systems - Reinforcement learning - Robotics - Search engine optimization - Social choice theory SUBMISSION INSTRUCTIONS Submitted papers should be in the ICML 2023 format (NOT ANONYMIZED) and up to 6 pages long, excluding references and appendix. Accepted papers will be presented as posters or contributed oral presentations. Submissions should be uploaded as a single pdf file at https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fopenreview.net%2Fgroup%3Fid%3DICML.cc%2F2023%2FWorkshop%2FMFPL&data=05%7C01%7Cuai%40engr.orst.edu%7C658116502d6c49af055f08db6b5d5415%7Cce6d05e13c5e4d6287a84c4a2713c113%7C0%7C0%7C638221820814333160%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=J3EqmbUq6goGbzJ4gHI9thY%2B0YKVHok%2BdVuRJdsvT7U%3D&reserved=0 CONFIRMED INVITED SPEAKERS Eytan Bakshy (Meta) Vincent Conitzer (CMU) Vineet Goyal (Columbia) Chi Jin (Princeton) Thorsten Joachims (Cornell) Sanmi Koyejo (Stanford) Dorsa Sadigh (Stanford) Yisong Yue (Caltech)
_______________________________________________ uai mailing list uai@engr.orst.edu https://it.engineering.oregonstate.edu/mailman/listinfo/uai
