Tag
This paper investigates whether LLM agents can infer hidden world models through interaction, finding that they struggle to build stable internal models as complexity increases.
A research paper proposes agentic automata learning to evaluate whether LLM agents can infer hidden world models through interaction, finding that performance drops sharply as task complexity increases and that reasoning models outperform non-reasoning ones but still struggle.
This tutorial provides a complete implementation of the TTT algorithm for active automata learning in Python, explaining how it improves over L* by eliminating redundant membership queries and combining discrimination trees with binary search counterexample analysis.
This paper presents a novel framework for synthesizing finite-state controllers for Partially Observable Markov Decision Processes (POMDPs) by integrating sampling, automata learning, and model-checking. The approach provides formal guarantees for threshold-safety problems that elude existing formal synthesis tools.
This paper introduces LANTERN, a framework for multi-source neurosymbolic transfer in reinforcement learning that uses LLMs to generate task automata and adaptive gating to improve sample efficiency.