Think Less, Know More: State-Aware Reasoning Compression with Knowledge Guidance for Efficient Reasoning

2026-04-10

AI Summary

This work addresses the challenge of efficiently eliminating redundancy and bias in large reasoning models without compromising accuracy. To this end, the authors propose STACK, a framework that explicitly models sources of redundancy during reasoning and employs a state-aware dynamic compression strategy: it leverages retrieval-augmented knowledge-guided compression under uncertain or biased states, switches to self-prompted compression in high-confidence verbose states, and incorporates an early-stopping mechanism based on answer convergence. By integrating online construction of short-long contrastive samples with joint PPO-DPO training, STACK achieves a 59.9% average reduction in response length while improving accuracy by 4.8 percentage points across three mathematical reasoning benchmarks, substantially enhancing the trade-off between efficiency and performance.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Large Reasoning Models (LRMs) achieve strong performance on complex tasks by leveraging long Chain-of-Thought (CoT), but often suffer from overthinking, leading to excessive reasoning steps and high inference latency. Existing CoT compression methods struggle to balance accuracy and efficiency, and lack fine-grained, step-level adaptation to redundancy and reasoning bias. Therefore, we propose State-Aware Reasoning Compression with Knowledge Guidance (STACK), a framework that performs step-wise CoT compression by explicitly modeling stage-specific redundancy sources and integrating with a retrieval-augmented guidance. STACK constructs online long-short contrastive samples and dynamically switches between knowledge-guided compression for uncertain or biased reasoning state and self-prompted compression for overly long but confident state, complemented by an answer-convergence-based early stopping mechanism to suppress redundant verification. We further propose a reward-difference-driven training strategy by combining Proximal Policy Optimization (PPO) and Direct Preference Optimization (DPO), enabling models to learn state-conditioned compression strategies. Experiments on three mathematical reasoning benchmarks show that STACK achieves a superior accuracy-efficiency balance, reducing average response length by 59.9% while improving accuracy by 4.8 points over existing methods.

Problem

• Chain-of-Thought Compression
• Reasoning Efficiency
• Overthinking
• Redundancy
• Reasoning Bias

Innovation

• reasoning compression
• state-aware adaptation
• knowledge-guided reasoning
• retrieval-augmented generation
• efficiency-accuracy trade-off