Zero-Shot Instruction Following in RL via Structured LTL Representations

TL;DR

This paper introduces a hierarchical neural approach that uses structured LTL representations to improve zero-shot instruction following in multi-task reinforcement learning, enabling better generalisation to unseen tasks.

Contribution

It proposes a novel method that conditions policies on Boolean formula sequences from automata, with an attention mechanism for reasoning about subgoals, enhancing generalisation.

Findings

Outperforms existing methods in complex environments

Demonstrates strong zero-shot generalisation to unseen tasks

Effectively captures logical and temporal task structures

Abstract

We study instruction following in multi-task reinforcement learning, where an agent must zero-shot execute novel tasks not seen during training. In this setting, linear temporal logic (LTL) has recently been adopted as a powerful framework for specifying structured, temporally extended tasks. While existing approaches successfully train generalist policies, they often struggle to effectively capture the rich logical and temporal structure inherent in LTL specifications. In this work, we address these concerns with a novel approach to learn structured task representations that facilitate training and generalisation. Our method conditions the policy on sequences of Boolean formulae constructed from a finite automaton of the task. We propose a hierarchical neural architecture to encode the logical structure of these formulae, and introduce an attention mechanism that enables the policy to reason about future subgoals. Experiments in a variety of complex environments demonstrate the strong generalisation capabilities and superior performance of our approach.