AI Seminar Series: Dustin Morrill, Efficient Deviation Types and Learning for Hindsight... (July 30)
Dustin Morrill presents "Efficient Deviation Types and Learning for Hindsight Rationality in Extensive-Form Games" at the AI Seminar (July 30, 2021).
The Artificial Intelligence (AI) Seminar is a weekly meeting at the University of Alberta where researchers interested in AI can share their research. Presenters include both local speakers from the University of Alberta and visitors from other institutions. Topics related in any way to artificial intelligence, from foundational theoretical work to innovative applications of AI techniques to new fields and problems, are explored.
Abstract: Hindsight rationality is an approach to playing general-sum games that prescribes no-regret learning dynamics for individual agents with respect to a set of deviations, and further describes jointly rational behavior among multiple agents with mediated equilibria. To develop hindsight rational learning in sequential decision-making settings, we formalize behavioral deviations as a general class of deviations that respect the structure of extensive-form games. Integrating the idea of time selection into counterfactual regret minimization (CFR), we introduce the extensive-form regret minimization (EFR) algorithm that achieves hindsight rationality for any given set of behavioral deviations with computation that scales closely with the complexity of the set. We identify behavioral deviation subsets, the partial sequence deviation types, that subsume previously studied types and lead to efficient EFR instances in games with moderate lengths. In addition, we present a thorough empirical analysis of EFR instantiated with different deviation types in benchmark games, where we find that stronger types typically induce better performance.
Bio: Dustin is a Ph.D. candidate at the University of Alberta and the Alberta Machine Intelligence Institute (Amii) working with Professor Michael Bowling. He works on multi-agent learning and scaleable, dependable learning algorithms. He is a coauthor of [DeepStack](https://www.deepstack.ai) and he created [Cepheus's public match interface](http://poker-play.srv.ualberta.ca/). He completed a B.Sc. and M.Sc. in computing science at the University of Alberta where his M.Sc was also supervised by Michael Bowling. As an undergraduate, he worked with the Computer Poker Research Group (CPRG) to create an [open-source web interface to play against poker bots](https://github.com/dmorrill10/acpc_poker_gui_client) and to develop the 1st-place 3-player Kuhn poker entry in the 2014 Annual Computer Poker Competition (ACPC).
Видео AI Seminar Series: Dustin Morrill, Efficient Deviation Types and Learning for Hindsight... (July 30) канала Amii
The Artificial Intelligence (AI) Seminar is a weekly meeting at the University of Alberta where researchers interested in AI can share their research. Presenters include both local speakers from the University of Alberta and visitors from other institutions. Topics related in any way to artificial intelligence, from foundational theoretical work to innovative applications of AI techniques to new fields and problems, are explored.
Abstract: Hindsight rationality is an approach to playing general-sum games that prescribes no-regret learning dynamics for individual agents with respect to a set of deviations, and further describes jointly rational behavior among multiple agents with mediated equilibria. To develop hindsight rational learning in sequential decision-making settings, we formalize behavioral deviations as a general class of deviations that respect the structure of extensive-form games. Integrating the idea of time selection into counterfactual regret minimization (CFR), we introduce the extensive-form regret minimization (EFR) algorithm that achieves hindsight rationality for any given set of behavioral deviations with computation that scales closely with the complexity of the set. We identify behavioral deviation subsets, the partial sequence deviation types, that subsume previously studied types and lead to efficient EFR instances in games with moderate lengths. In addition, we present a thorough empirical analysis of EFR instantiated with different deviation types in benchmark games, where we find that stronger types typically induce better performance.
Bio: Dustin is a Ph.D. candidate at the University of Alberta and the Alberta Machine Intelligence Institute (Amii) working with Professor Michael Bowling. He works on multi-agent learning and scaleable, dependable learning algorithms. He is a coauthor of [DeepStack](https://www.deepstack.ai) and he created [Cepheus's public match interface](http://poker-play.srv.ualberta.ca/). He completed a B.Sc. and M.Sc. in computing science at the University of Alberta where his M.Sc was also supervised by Michael Bowling. As an undergraduate, he worked with the Computer Poker Research Group (CPRG) to create an [open-source web interface to play against poker bots](https://github.com/dmorrill10/acpc_poker_gui_client) and to develop the 1st-place 3-player Kuhn poker entry in the 2014 Annual Computer Poker Competition (ACPC).
Видео AI Seminar Series: Dustin Morrill, Efficient Deviation Types and Learning for Hindsight... (July 30) канала Amii
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![DLRLSS 2019 - Bandits - Csaba Szepesvári](https://i.ytimg.com/vi/eJ3wZ14RRBQ/default.jpg)
![The Tea Time Talks: Junfeng Wen, Batch Stationary Distribution Estimation (June 24)](https://i.ytimg.com/vi/J00ZX21tKTk/default.jpg)
![Lili Mou Machine Learning Course - Class 14: Softmax Regression](https://i.ytimg.com/vi/918iFgGTAwE/default.jpg)
![TechAid Presents TechTalks: Privacy & Fairness featuring Nidhi Hegde](https://i.ytimg.com/vi/t8OIuY0eMBg/default.jpg)
![The Tea Time Talks: Matthew E. Taylor, Assisting RL with External Information (June 16)](https://i.ytimg.com/vi/FjKEiPUF-Qs/default.jpg)
![The Tea Time Talks with Bo Liu (July 19, 2018)](https://i.ytimg.com/vi/mSM2fsMhOsM/default.jpg)
![Amii's Monthly AI Meetup - June 2020](https://i.ytimg.com/vi/SytkiBSwoXs/default.jpg)
![AI Seminar 2021: Chenjun Xiao, "On the Optimality of Batch Policy Optimization Algorithms"](https://i.ytimg.com/vi/DLav2FLK_nU/default.jpg)
![StreamML: Breaking down barriers to machine learning](https://i.ytimg.com/vi/T8cc-NVwhJY/default.jpg)
![Upper Bound 2023: Artificial Intelligence; Authentic Identity?](https://i.ytimg.com/vi/li4j0ISzPPs/default.jpg)
![TechAid Presents DevCon 2022: Create Magic With Data (Visualization) and d3.js](https://i.ytimg.com/vi/zTok_pvZ-q4/default.jpg)
![Meet 10 organizations using AI to transform health and wellness](https://i.ytimg.com/vi/ubwo8x2LJP0/default.jpg)
![AI in Games: Kevin Leyton-Brown, Machine Learning for Optimization and Game Theory](https://i.ytimg.com/vi/Qu7nhGbRdAg/default.jpg)
![AI in Games: Matthew Guzdial, Computational Creativity](https://i.ytimg.com/vi/j3YzRvo1GTc/default.jpg)
![Upper Bound 2023: 5 Years of GPTs with Finbarr Timbers](https://i.ytimg.com/vi/t05TdSdoZPo/default.jpg)
![AI Seminar: Amir Hossein - "GPEX, A Framework For Interpreting Artificial Neural Networks"](https://i.ytimg.com/vi/08zjKrrGosU/default.jpg)
![AI Seminar: Matt Ginsberg - What I Learned from Dr. Fill](https://i.ytimg.com/vi/ZKJKEkpVuf4/default.jpg)
![Tea Time Talks: Rupam Mahmood, New Forms of Policy Gradients for Model-free Estimation (June 11)](https://i.ytimg.com/vi/8kNAhJNWsig/default.jpg)
![The Tea Time Talks: Martin Müller, Some Old Ideas about Search and Learning (June 17)](https://i.ytimg.com/vi/DIHUWFdLUec/default.jpg)
![AI Seminar Series: Sriram Ganapathi Subramanian, Recent Advances in Mean-Field RL Methods (May 14)](https://i.ytimg.com/vi/GcGIZMOkB6Y/default.jpg)
![JUNO Talks: Matching the Music to the Mood for Your Role Playing Game](https://i.ytimg.com/vi/2jTh1GHlkqw/default.jpg)