DRIFT：用于高效MLLM微调的推理先验转移
DRIFT: Transferring Reasoning Priors for Efficient MLLM Fine-Tuning
AI总结：
DRIFT通过在梯度空间中转移推理知识，实现高效稳定的多模态推理迁移，优于传统方法并在数据和计算上更高效。
摘要：
多模态大语言模型（MLLMs）虽进展迅速，但推理能力常逊于强文本模型。通常需大规模多模态推理数据或强化学习，成本高昂。参数空间模型融合是替代方案，但效果不稳定。DRIFT是一种轻量方法，通过梯度空间转移推理知识并保持多模态对齐。DRIFT预计算文本推理专家与多模态模型参数差异的推理先验，并在监督微调中偏指导梯度。该设计保留了标准SFT流程的简洁性，同时实现高效稳定的推理迁移。实验表明，DRIFT在MathVista和MathVerse等多模态推理基准上，优于传统方法并在数据和计算上更高效。
英文摘要：
Multimodal large language models (MLLMs) have made rapid progress, yet their reasoning ability often lags behind strong text-only LLMs. Bridging this gap typically requires large-scale multimodal reasoning data or reinforcement learning, incurring substantial cost. An appealing alternative is parameter-space model merging between reasoning-enhanced LLMs and MLLMs, but we show that naive merging is fragile: its effectiveness varies widely across model families and can significantly degrade performance (e.g., for Qwen-based MLLMs). We propose Directional Reasoning Injection for Fine-Tuning (DRIFT), a lightweight method that transfers reasoning knowledge in the gradient space while preserving multimodal alignment. DRIFT precomputes a reasoning prior from the parameter differences between text-only reasoning experts and multimodal models, and uses it to bias gradients during supervised fine-tuning. This design retains the simplicity of standard SFT pipelines while enabling efficient and stable reasoning transfer. Experiments on multimodal reasoning benchmarks, including MathVista and MathVerse, show that DRIFT consistently outperforms naive merging and standard SFT, and matches or surpasses training-intensive methods with substantially lower data and compute.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
