cs.LGFeb 15, 2026

Zero-Shot Instruction Following in RL via Structured LTL Representations

Authors: Mathias Jackermeier, Mattia Giuri, Jacques Cloete, Alessandro Abate

Organizations: Department of Computer Science, University of Oxford · Oxford Robotics Institute, University of Oxford

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.

Figures & tables

Appendix figures & tables10 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. PlatoLTL: Scaling LTL-Guided Multi-Task RL

    Jan 30, 2026Jacques Cloete, Mathias Jackermeier, Ioannis Havoutis +1RL for RoboticsConstrained RL

  2. Temporal Logic Guided Universal Task Representations for Reinforcement Learning

    Aug 16, 2026Hao Zhang, Zhangli Zhou, Zhen KanRepresentation Learning for RLReinforcement Learning

  3. Jaxolotl: A Unified High-Performance Benchmark Suite for LTL-Based Multi-Task RL

    Sep 29, 2026Mathias Jackermeier, Jacques Cloete, Alessandro AbateRL BenchmarksReinforcement Learning