OPE: Overcoming Information Saturation in Parallel Thinking via Outline-Guided Path Exploration

TL;DR

This paper introduces OPE, a novel method that enhances parallel reasoning in large models by partitioning solution space with outlines, reducing redundancy, and improving problem-solving effectiveness.

Contribution

It proposes Outline-Guided Path Exploration (OPE), a new approach that explicitly partitions solution space to improve diversity and performance in parallel reasoning models.

Findings

OPE improves reasoning accuracy across multiple benchmarks.

OPE enhances diversity of exploration paths, reducing redundancy.

OPE outperforms existing methods in large reasoning models.

Abstract

Parallel thinking has emerged as a new paradigm for large reasoning models (LRMs) in tackling complex problems. Recent methods leverage Reinforcement Learning (RL) to enhance parallel thinking, aiming to address the limitations in computational resources and effectiveness encountered with supervised fine-tuning. However, most existing studies primarily focus on optimizing the aggregation phase, with limited attention to the path exploration stage. In this paper, we theoretically analyze the optimization of parallel thinking under the Reinforcement Learning with Verifiable Rewards (RLVR) setting, and identify that the mutual information bottleneck among exploration paths fundamentally restricts overall performance. To address this, we propose Outline-Guided Path Exploration (OPE), which explicitly partitions the solution space by generating diverse reasoning outlines prior to parallel path reasoning, thereby reducing information redundancy and improving the diversity of information captured across exploration paths. We implement OPE with an iterative RL strategy that optimizes outline planning and outline-guided reasoning independently. Extensive experiments across multiple challenging mathematical benchmarks demonstrate that OPE effectively improves reasoning performance in different aggregation strategies, enabling LRMs to more reliably discover correct solutions.