Coupled autoregressive active inference agents for control of multi-joint dynamical systems

TL;DR

This paper introduces a coupled active inference agent composed of multiple autoregressive subagents that collaboratively learn and control multi-joint dynamical systems, effectively balancing exploration and goal achievement.

Contribution

It presents a novel coupled autoregressive active inference framework that improves learning and control of complex mechanical systems over uncoupled agents.

Findings

Successfully learned the dynamics of a double mass-spring-damper system

Achieved control to drive the system to a desired position

Outperformed uncoupled subagents in surprise and goal alignment

Abstract

We propose an active inference agent to identify and control a mechanical system with multiple bodies connected by joints. This agent is constructed from multiple scalar autoregressive model-based agents, coupled together by virtue of sharing memories. Each subagent infers parameters through Bayesian filtering and controls by minimizing expected free energy over a finite time horizon. We demonstrate that a coupled agent of this kind is able to learn the dynamics of a double mass-spring-damper system, and drive it to a desired position through a balance of explorative and exploitative actions. It outperforms the uncoupled subagents in terms of surprise and goal alignment.