General Dynamic Goal Recognition using Goal-Conditioned and Meta Reinforcement Learning

TL;DR

This paper introduces the GDGR problem, a new framework for real-time goal recognition in dynamic environments, and proposes two reinforcement learning-based methods, GC-AURA and Meta-AURA, for rapid adaptation and accurate recognition.

Contribution

The paper presents two novel reinforcement learning approaches, GC-AURA and Meta-AURA, for dynamic goal recognition in changing environments, extending existing methods to real-time, adaptive scenarios.

Findings

GC-AURA generalizes to new goals effectively.

Meta-AURA adapts to new environments with high accuracy.

Both methods perform well under noisy conditions.

Abstract

Understanding an agent's goal through its behavior is a common AI problem called Goal Recognition (GR). This task becomes particularly challenging in dynamic environments where goals are numerous and ever-changing. We introduce the General Dynamic Goal Recognition (GDGR) problem, a broader definition of GR aimed at real-time adaptation of GR systems. This paper presents two novel approaches to tackle GDGR: (1) GC-AURA, generalizing to new goals using Model-Free Goal-Conditioned Reinforcement Learning, and (2) Meta-AURA, adapting to novel environments with Meta-Reinforcement Learning. We evaluate these methods across diverse environments, demonstrating their ability to achieve rapid adaptation and high GR accuracy under dynamic and noisy conditions. This work is a significant step forward in enabling GR in dynamic and unpredictable real-world environments.