ViCuR: 视觉线索作为多模态在策略蒸馏中的可恢复特权
ViCuR: Visual Cues as Recoverable Privilege for Multimodal On-Policy Distillation
AI总结：
提出ViCuR框架，通过将教师特权从答案侧替换为输入中的视觉线索，并引入轻量级线索恢复模块，解决多模态在策略蒸馏中的训练-测试不匹配问题，在七个基准上显著提升学生模型性能。
摘要：
在策略蒸馏（OPD）通过在教师监督下，对学生自身策略采样的轨迹进行训练来改进推理。在多模态推理中，一种常见的扩展是使用特权教师，该教师观察仅在训练时可用的信号，如参考答案或理由。然而，这种答案侧特权造成了训练-测试不匹配：教师的监督可能依赖于学生无法获得的信号，鼓励捷径模仿而非基于视觉的推理。我们提出ViCuR，一种基于视觉的特权教师蒸馏框架，用视觉线索（输入中与查询相关的证据）取代答案侧特权。由于这些线索来源于推理时可用的相同视觉输入，它们的证据可由学生恢复。为此，ViCuR引入了一个轻量级线索恢复模块，在预填充期间使用专用的汇点令牌交叉注意力，将任务相关的视觉证据聚合到内部表示中，而不改变推理接口或需要辅助的线索生成损失。在七个基准上，使用Qwen3-VL-2B和8B学生，ViCuR在总体平均性能上持续优于基于答案的在策略自蒸馏，分别提升+1.19和+1.24。它还能自然地扩展到更强的教师OPD，超越OPD基线+0.64和+1.08，并在8B规模上具有一致的域外增益。这些结果表明，在多模态在策略蒸馏中，教师特权的设计与教师强度同等重要。
英文摘要：
On-policy distillation (OPD) improves reasoning by training a student on trajectories sampled from its own policy under supervision from a teacher. In multimodal reasoning, a common extension is to use a privileged teacher that observes training-time-only signals such as reference answers or rationales. However, such answer-side privilege creates a train-test mismatch: the teacher's supervision may depend on signals unavailable to the student, encouraging shortcut imitation rather than visually grounded reasoning. We propose ViCuR, a visually grounded privileged-teacher distillation framework that replaces answer-side privilege with visual cues (query-related evidence in the input). Because these cues are derived from the same visual input available at inference, their evidence is recoverable by the student. To support this, ViCuR introduces a lightweight cue recovery module that uses dedicated sink-token cross-attention during prefill to aggregate task-relevant visual evidence into an internal representation, without changing the inference interface or requiring auxiliary cue-generation losses. Across seven benchmarks with Qwen3-VL-2B and 8B students, ViCuR consistently improves over answer-based on-policy self-distillation by +1.19 and +1.24 on overall average performance. It also extends naturally to stronger-teacher OPD, surpassing OPD baselines by +0.64 and +1.08, with consistent out-of-domain gains at the 8B scale. These results show that, in multimodal on-policy distillation, the design of teacher privilege is as important as teacher strength.
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。
