Saved in:
Bibliographic Details
Main Authors: Song, Zeen, Ma, Zihao, Qiang, Wenwen, Zheng, Changwen, Hua, Gang
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.06493
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911427491528704
author Song, Zeen
Ma, Zihao
Qiang, Wenwen
Zheng, Changwen
Hua, Gang
author_facet Song, Zeen
Ma, Zihao
Qiang, Wenwen
Zheng, Changwen
Hua, Gang
contents Inference-time reasoning scaling has significantly advanced the capabilities of Large Language Models (LLMs) in complex problem-solving. A prevalent approach involves external search guided by Process Reward Models (PRMs). However, a fundamental limitation of this framework is the epistemic uncertainty of PRMs when evaluating reasoning paths that deviate from their training distribution. In this work, we conduct a systematic analysis of this challenge. We first provide empirical evidence that PRMs exhibit high uncertainty and unreliable scoring on out-of-distribution (OOD) samples. We then establish a theoretical framework proving that while standard search incurs linear regret accumulation, an uncertainty-aware strategy can achieve sublinear regret. Motivated by these findings, we propose Uncertainty-Aware Tree Search (UATS), a unified method that estimates uncertainty via Monte Carlo Dropout and dynamically allocates compute budget using a reinforcement learning-based controller. Extensive experiments demonstrate that our approach effectively mitigates the impact of OOD errors.
format Preprint
id arxiv_https___arxiv_org_abs_2602_06493
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Adaptive Uncertainty-Aware Tree Search for Robust Reasoning
Song, Zeen
Ma, Zihao
Qiang, Wenwen
Zheng, Changwen
Hua, Gang
Machine Learning
Inference-time reasoning scaling has significantly advanced the capabilities of Large Language Models (LLMs) in complex problem-solving. A prevalent approach involves external search guided by Process Reward Models (PRMs). However, a fundamental limitation of this framework is the epistemic uncertainty of PRMs when evaluating reasoning paths that deviate from their training distribution. In this work, we conduct a systematic analysis of this challenge. We first provide empirical evidence that PRMs exhibit high uncertainty and unreliable scoring on out-of-distribution (OOD) samples. We then establish a theoretical framework proving that while standard search incurs linear regret accumulation, an uncertainty-aware strategy can achieve sublinear regret. Motivated by these findings, we propose Uncertainty-Aware Tree Search (UATS), a unified method that estimates uncertainty via Monte Carlo Dropout and dynamically allocates compute budget using a reinforcement learning-based controller. Extensive experiments demonstrate that our approach effectively mitigates the impact of OOD errors.
title Adaptive Uncertainty-Aware Tree Search for Robust Reasoning
topic Machine Learning
url https://arxiv.org/abs/2602.06493