cs.CVNov 18, 2026

TALD: Transferability-Aware Logit Disentanglement for Cross-Modal Knowledge Distillation

Authors: Hui Li, Pengfei Yang, Xinyi Feng, Junhong Liu, Tianyang Zheng, Hui Zeng, Le Dong, Kaibing Zhang, 

Organizations: 1Xidian University · 2Beihang University · 3Xi’an Polytechnic University

Abstract

Cross-modal knowledge distillation can incur negative transfer when source-modality supervision conflicts with the target modality's discriminative characteristics. We propose TALD, a two-stage framework for Transferability-Aware Logit Disentanglement in class-aligned logit space. Guided by task-dependent self-benefiting and cross-modal exclusivity criteria, a Residual Disentangler decomposes logits from a teacher and an independently trained target-modality reference student into modality-general and modality-specific components. A Sample-Adaptive Modulator regulates decomposition constraints and sample-specific fusion weights. The learned modality-general components are then adaptively fused to supervise the final student. Across five benchmarks spanning vision, audio, and text, TALD achieves the largest mean accuracy gain and highest positive-transfer frequency among the compared methods. Across ten transfer directions with five runs each, TALD improves over training without distillation in 84% of evaluations, compared with at most 70% for the baselines, while offering a favorable accuracy-efficiency trade-off. We plan to release the code.

Figures & tables

Appendix figures & tables15 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Cross-Modal Knowledge Distillation without Paired Data: Theoretical Foundation and Algorithm

    Jun 9, 2026Trong Khiem Tran, Anh Duc Chu, Quang Hung Pham +2Cross-Modal Knowledge DistillationCross-Modal Alignment

  2. Codebook-Guided Cross-Modal Knowledge Distillation for Structurally Heterogeneous Features

    Sep 29, 2026Dae Ung Jo, Jongin Lim, YoungJoon Yoo +1Cross-Modal Knowledge DistillationCross-Modal Alignment

  3. From Images to Words: Efficient Cross-Modal Knowledge Distillation to Language Models from Black-box Teachers

    Mar 11, 2026Ayan Sengupta, Shantanu Dixit, Md Shad Akhtar +1Black-Box Knowledge DistillationCross-Modal Knowledge Distillation