Problem
Research questions and friction points this paper is trying to address.
multimodal reasoning
computational costs
KV-cache pressure
cross-modal constraints
visual laziness
MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought Compression
๐ค AI Summary
ๆฌๆๆๅบMCPOๆนๆณ๏ผ้่ฟไธค้ถๆฎตๅ็ผฉๅๅฏน้ฝไผๅๅคๆจกๆ้ฟ้พๆจ็๏ผๅๅฐ่ฎก็ฎๆๆฌๅๅไฝ๏ผๆ้ซๆจ็ๆ็ใ
๐ Abstract
Recently, multimodal large-scale reasoning models have demonstrated remarkable capabilities in solving complex tasks through long Chains-of-Thought (M-CoT). However, excessively long reasoning trajectories incur substantial computational costs and significant KV-cache pressure. Existing CoT compression and alignment paradigms mainly rely on static rules or single-dimensional preferences, lacking fine-grained cross-modal constraints; as a result, they are prone to inducing visual laziness and hallucinatory reasoning. To address these issues, we propose Modality-Contrastive Preference Optimization (MCPO), a highly sample-efficient two-stage length-compression method that requires fewer than 900 training samples. In the compression stage, we introduce a step-level Normalized Cross-Modal Mutual Information (NCMI) pruning algorithm, which automatically identifies and removes visual-independent reasoning steps by comparing the reasoning discrepancies between with-image and no-image contexts. This significantly reduces redundancy and hallucinatory content in the reasoning chains. In the alignment stage, the model first undergoes supervised fine-tuning to achieve domain-adaptive initialization, followed by optimization using an asymmetric multimodal length-controlled preference loss. This objective adopts a highly nonlinear odds-ratio formulation that provides steep gradients in the with-image context to reinforce length constraints for preferred trajectories, while applying a scaled, flat-gradient linear difference in the no-image context to maintain modality consistency, thereby achieving stable cross-modal preference alignment. Extensive experiments on mainstream base models such as Qwen3-VL-Thinking show that our method can reduce CoT length by up to 69.5% and achieve up to 3.34x end-to-end inference speedup while preserving original accuracy.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Modality-Contrastive Preference Optimization (MCPO)
Normalized Cross-Modal Mutual Information (NCMI) pruning
asymmetric multimodal length-controlled preference loss
Chain-of-Thought (CoT) compression
multimodal reasoning
๐ Similar Papers
No similar papers found.
