Jul 2026
CoCaRS: Correlation Calibration-Based Redundancy Suppression for Heterogeneous Knowledge Distillation
Correlation Calibration-based Redundancy Suppression (CoCaRS) is proposed to better retain structural information while suppressing redundancy and reduce sensitivity to coefficient settings across teacher-student pairs and training stages.
Feng Yu, Haiwei Pan, Kejia Zhang et al.
· arXiv.org · 0 citations