Loading Events

« All Events

  • This event has passed.

Multi-Agent Learning in Dynamic Environments Workshop

April 26 @ 9:00 am 5:00 pm

Blocker 128

Abstract: Multi-agent learning and decision-making in networked and dynamic environments are at the forefront of challenges that artificial intelligence faces. That is, how can a team of learners coordinate to do better when there are limitations and hurdles to their interactions in the form of privacy, communication, or security? How can agents learn, adapt, and decide in the presence of other agents who strive to make decisions to further their own objectives? This workshop on “Multi-Agent Learning in Dynamic Environments” aims to provide a landscape of recent research activities on multi-agent learning in dynamic environments, and foster potential collaborations among the researchers broadly working in the fields of reinforcement learning, game theory, optimization and control, and those that are interested in engineering and science applications of such methods. This workshop will consist of seminar-style talks followed by a discussion, and related short presentations. 

Speakers

Safe Learning for Dynamical Systems and Control

Santiago Paternain
Assistant Professor, Department of Electrical, Computer and Systems Engineering
Rensselaer Polytechnic Institute

The Role of Lookahead in Reinforcement Learning Algorithms

Anna Winnicki
PhD Candidate, Department of Electrical and Computer Engineering
University of Illinois Urbana-Champaign

Multi-Agent Reinforcement Learning for Large-Scale Markov Potential Games

Dongsheng Ding
Postdoctoral Researcher, Department of Electrical and Systems Engineering
University of Pennsylvania

Schedule

TimePresenterTopic
8:50 AMCeyhun EksinWelcome and Introductions
9:00 AM–9:50 AMSantiago PaternainSafe Learning for Dynamical Systems and Control
9:50 AM–10:00 AMBreak
10:00 AM–10:30 AMResearch Highlights
Soham Das, Ceyhun EksinLearning Nash in Constrained Markov GAmes with an α-Potential
Steve Suh, Bin WuTrust-Driven Collaboration: Mechanisms to Facilitate Multi-Robot Cooperation in Constrained Environments
Ujwal Dinesha, Srinivas ShakkotaiA Multi-Agent View of Wireless Video Streaming with Delayed Client-Feedback
10:30 AM–11:20 AMAnna WinnickiThe Role of Lookahead in Reinforcement Learning Algorithms
11:20 AM–11:30 AMBreak
11:30 AM–12:00 PMResearch Highlights
Khaled Nakhleh, Ceyhun EksinSimulation-Based Optimistic Policy Iteration in Multi-Agent Games with Kullback-Leibler Control Cost
Jiachen Xi, Alfredo Garcia, Petar MomcilovicMulti-Agent Reinforcement Learning for Multi-Area Power Exchange
Austin Carroll, Bharath Sivaram, Swaminathan Goppalswamy, Lucas KrakowCooperative Target Tracking: What did you see?
12:00 PM–2:00 PMLunch
2:00 PM–2:50 PMDongsheng DingMulti-Agent Reinforcement Learning for Large-Scale Markov Potential Games
2:50 PM–3:00 PMBreak
3:00 PM–3:30 PMResearch Highlights
Ruida Zhou, Chao TianProvable Policy Gradient Methods for Average-Reward Markov Potential Games
Zhide Wang, Yanling ChangStructural Estimation of Partially Observable Dynamic Games: an Application of Retail Investment