用图文双教师提升小模型性能，效果最高提升4.49%。
Powerful Teachers Matter: Text-Guided Multi-view Knowledge Distillation with Visual Prior Enhancement
- 图文双教师协同：视觉+文本（CLIP）提供多模态监督信号。
- 增强视觉教师：引入边缘与高频特征，提升细节感知能力。
- 适合追求高精度的轻量化模型部署者，尤其关注视觉语义对齐。
知识蒸馏将大模型知识迁移到小模型以实现高效推理。现有方法多关注蒸馏策略，却忽视教师模型知识质量的提升。本文提出文本引导的多视角知识蒸馏（TMKD），采用双模态教师——视觉教师与文本教师（CLIP），提供更丰富的监督信号。具体地，通过融合边缘与高频特征等视觉先验，增强视觉教师的多视角输入；文本教师则通过感知先验的提示生成语义权重，指导自适应特征融合。此外，引入视觉-语言对比正则化，强化学生模型的语义知识。在五个基准上的大量实验表明，TMKD性能最高提升4.49%，验证了双教师多视角增强策略的有效性。代码已公开于https://anonymous.4open.science/r/TMKD-main-44D1。
原文摘要 · Abstract (English)
Knowledge distillation transfers knowledge from large teacher models to smaller students for efficient inference. While existing methods primarily focus on distillation strategies, they often overlook the importance of enhancing teacher knowledge quality. In this paper, we propose Text-guided Multi-view Knowledge Distillation (TMKD), which leverages dual-modality teachers, a visual teacher and a text teacher (CLIP), to provide richer supervisory signals. Specifically, we enhance the visual teacher with multi-view inputs incorporating visual priors (edge and high-frequency features), while the text teacher generates semantic weights through prior-aware prompts to guide adaptive feature fusion. Additionally, we introduce vision-language contrastive regularization to strengthen semantic knowledge in the student model. Extensive experiments on five benchmarks demonstrate that TMKD consistently improves knowledge distillation performance by up to 4.49\%, validating the effectiveness of our dual-teacher multi-view enhancement strategy. Code is available at https://anonymous.4open.science/r/TMKD-main-44D1.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。