CVAug 6, 2025

TopKD: Top-scaled Knowledge Distillation

arXiv:2508.04539v11 citationsh-index: 1
Originality Incremental advance
AI Analysis

This addresses the challenge of efficient model compression for machine learning practitioners, though it is incremental as it builds on existing logit-based distillation methods.

The paper tackles the problem of knowledge distillation by focusing on Top-K knowledge in teacher logit distributions, proposing TopKD, which consistently surpasses state-of-the-art methods on datasets like CIFAR-100 and ImageNet.

Recent advances in knowledge distillation (KD) predominantly emphasize feature-level knowledge transfer, frequently overlooking critical information embedded within the teacher's logit distributions. In this paper, we revisit logit-based distillation and reveal an underexplored yet critical element: Top-K knowledge. Motivated by this insight, we propose Top-scaled Knowledge Distillation (TopKD), a simple, efficient, and architecture-agnostic framework that significantly enhances logit-based distillation. TopKD consists of two main components: (1) a Top-K Scaling Module (TSM), which adaptively amplifies the most informative logits, and (2) a Top-K Decoupled Loss (TDL), which offers targeted and effective supervision. Notably, TopKD integrates seamlessly into existing KD methods without introducing extra modules or requiring architectural changes. Extensive experiments on CIFAR-100, ImageNet, STL-10, and Tiny-ImageNet demonstrate that TopKD consistently surpasses state-of-the-art distillation methods. Moreover, our method demonstrates substantial effectiveness when distilling Vision Transformers, underscoring its versatility across diverse network architectures. These findings highlight the significant potential of logits to advance knowledge distillation.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes