CS Seminar: Jayesh Gupta, "Deep Implicit Coordination Graphs for Multi-agent Reinforcement Learning"

Wednesday, October 21, 2020
1:30 p.m.

Computer Science Reinforcement Learning Seminar Series (RLSS)

Deep Implicit Coordination Graphs for Multi-Agent Reinforcement Learning

Jayesh Gupta
Stanford and Microsoft Research

Multi-agent reinforcement learning (MARL) requires coordination to efficiently solve certain tasks. Fully centralized control is often infeasible in such domains due to the size of joint action spaces. Coordination graph based formalization allows reasoning about the joint action based on the structure of interactions. However, they often require domain expertise in their design. In this talk, we will discuss the recently introduced deep implicit coordination graph (DICG) architecture for such scenarios. DICG consists of a module for inferring the dynamic coordination graph structure which is then used by a graph neural network based module to learn to implicitly reason about the joint actions or values. DICG allows learning the tradeoff between full centralization and decentralization via standard actor-critic methods to significantly improve coordination for domains with large numbers of agents.

PRIMARY PAPER: https://arxiv.org/abs/2006.11438

All talks are listed at https://www.cs.umd.edu/talks/rlss, and recordings and slides will be linked there a few days after the talk.

remind we with google calendar


October 2020

27 28 29 30 1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31
1 2 3 4 5 6 7
Submit an Event