Hide to See: 面向VLM蒸馏中视觉锚定思维的推理前缀掩码
Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation
AI总结：
提出一种推理前缀掩码蒸馏框架，通过掩码学生模型的显著推理前缀，迫使其在推理过程中更依赖视觉证据，从而缓解长推理轨迹中的视觉遗忘问题，提升多模态推理性能。
摘要：
近期VLM中的思考-回答方法（如Qwen3-VL-Thinking）通过在最终答案前利用中间推理步骤来提升推理性能，但其计算成本显著增加，尤其是对于较大的VLM。为了将这种能力蒸馏到紧凑的思考-回答VLM中，一个主要目标是提高学生在整个推理轨迹中利用视觉证据的能力，因为长思考-回答轨迹存在视觉遗忘问题。为此，我们引入了一种新颖的思考-回答蒸馏框架，通过掩码学生模型的显著推理前缀，鼓励学生将思考锚定在视觉信息上。为了补偿这种被掩码的文本线索，学生在蒸馏过程中被鼓励更多地依赖视觉证据作为替代信息源。我们的掩码策略包括：1）逐token的显著推理前缀掩码，针对每个下一token预测选择性掩码高影响力的推理前缀；2）自调节掩码预算调度，根据教师-学生分布之间的差异（即蒸馏难度）逐渐增加掩码规模。在蒸馏阶段，学生模型由我们的显著推理前缀掩码引导，该掩码同时阻塞未来token和显著推理线索，替代了自回归语言建模中使用的标准因果掩码。实验结果表明，我们的方法在多模态推理基准上优于最近的开源VLM、VLM蒸馏和自蒸馏方法，进一步分析证实了学生思考过程中视觉利用的增强。
英文摘要：
Recent think-answer approaches in VLMs, such as Qwen3-VL-Thinking, boost reasoning performance by leveraging intermediate thinking steps before the final answer, but their computational cost becomes substantial, especially for larger VLMs. To distill such capabilities into compact think-answer VLMs, a primary objective is to improve the student's ability to utilize visual evidence throughout its reasoning trace, as long think-answer traces suffer from visual forgetting issues. To this end, we introduce a novel think-answer distillation framework that encourages the student to anchor its thinking on visual information by masking the student's salient reasoning prefixes. To compensate for such masked textual cues, the student is encouraged to rely more on visual evidence as an alternative source of information during distillation. Our masking strategies include: 1) token-wise salient reasoning-prefix masking, which masks high-influence reasoning prefixes selectively for each next-token prediction, and 2) self-paced masking budget scheduling, which gradually increases the masking scale according to distillation difficulty, measured by the discrepancy between teacher--student distributions. In the distillation phase, the student is guided by our salient reasoning-prefix mask, which blocks both future tokens and salient reasoning cues, in place of the standard causal mask used for auto-regressive language modeling. Experimental results show that our approach outperforms recent open-source VLMs, VLM distillation, and self-distillation methods on multimodal reasoning benchmarks, while further analyzes confirm enhanced visual utilization along the student thinking process.
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。
