Model based Multi-agent Reinforcement Learning with Tensor Decompositions

by   Pascal Van Der Vaart, et al.
University of Oxford
Delft University of Technology

A challenge in multi-agent reinforcement learning is to be able to generalize over intractable state-action spaces. Inspired from Tesseract [Mahajan et al., 2021], this position paper investigates generalisation in state-action space over unexplored state-action pairs by modelling the transition and reward functions as tensors of low CP-rank. Initial experiments on synthetic MDPs show that using tensor decompositions in a model-based reinforcement learning algorithm can lead to much faster convergence if the true transition and reward functions are indeed of low rank.


Reinforcement Learning in Factored Action Spaces using Tensor Decompositions

We present an extended abstract for the previously published work TESSER...

Tesseract: Tensorised Actors for Multi-Agent Reinforcement Learning

Reinforcement Learning in large action spaces is a challenging problem. ...

Mean-Field Approximation of Cooperative Constrained Multi-Agent Reinforcement Learning (CMARL)

Mean-Field Control (MFC) has recently been proven to be a scalable tool ...

Deceptive Reinforcement Learning in Model-Free Domains

This paper investigates deceptive reinforcement learning for privacy pre...

Tensor Decomposition for Multi-agent Predictive State Representation

Predictive state representation (PSR) uses a vector of action-observatio...

Scalable Multi-Agent Reinforcement Learning for Networked Systems with Average Reward

It has long been recognized that multi-agent reinforcement learning (MAR...

Learning Fair Canonical Polyadical Decompositions using a Kernel Independence Criterion

This work proposes to learn fair low-rank tensor decompositions by regul...

Please sign up or login with your details

Forgot password? Click here to reset