Thermodynamic Focusing for Inference-Time Search: Practical Methods for Target-Conditioned Sampling and Prompted Inference
This addresses a practical problem in language generation, planning, and reinforcement learning, offering incremental improvements through a hybrid approach.
The paper tackles the challenge of finding rare solutions in large search spaces by introducing the Inverted Causality Focusing Algorithm (ICFA), a framework for target-conditioned reweighting that adaptively controls focusing strength to avoid degeneracy, with experiments in constrained language generation and sparse-reward navigation.
Finding rare but useful solutions in very large candidate spaces is a recurring practical challenge across language generation, planning, and reinforcement learning. We present a practical framework, \emph{Inverted Causality Focusing Algorithm} (ICFA), that treats search as a target-conditioned reweighting process. ICFA reuses an available proposal sampler and a task-specific similarity function to form a focused sampling distribution, while adaptively controlling focusing strength to avoid degeneracy. We provide a clear recipe, a stability diagnostic based on effective sample size, a compact theoretical sketch explaining when ICFA can reduce sample needs, and two reproducible experiments: constrained language generation and sparse-reward navigation. We further show how structured prompts instantiate an approximate, language-level form of ICFA and describe a hybrid architecture combining prompted inference with algorithmic reweighting.