analiza aiqbench
This research advances zero-shot instruction following in multi-task reinforcement learning by leveraging Linear Temporal Logic (LTL) to encode complex, temporally extended tasks. The authors introduce a hierarchical neural architecture with an attention mechanism that conditions policies on Boolean formulae derived from finite automata, enabling agents to reason about future subgoals without prior training on specific tasks. Experimental results in diverse environments demonstrate superior generalization capabilities compared to existing approaches that fail to capture the full logical structure of LTL specifications.
Zero-Shot Instruction Following in RL via Structured LTL Representations
arXiv:2602.14344v2 Announce Type: replace 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.
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.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2602.14344 [cs.LG]
(or arXiv:2602.14344v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2602.14344 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 244 cuvinte
Subiecte conexe
Links #5: 2026/07
LessWrong · 31 iul. 2026
The Entangled Dimensions of Decision Theory
LessWrong · 30 iul. 2026
FireRedAudio: A General-Purpose Audio Language Model with Decoupled Continuous Representations for Understanding and Generation
arXiv · 26 aug. 2026
The J-Space Debate, Agent Swarms, and Pacing Frontier AI - Digital Minds Newsletter #4
LessWrong · 18 sept. 2026