paper-with-me

홈 › Papers

A Functional Perspective on Knowledge Distillation in Neural Networks

2025-10-14 · Israel Mason-Williams, Gabryel Mason-Williams, Helen Yannakoudakis arxiv

Knowledge distillation is considered a compression mechanism when judged on the resulting student's accuracy and loss, yet its functional impact is poorly understood. We quantify the compression capacity of knowledge distillation and the resulting knowledge transfer from a functional perspective, decoupling compression from architectural reduction to provide an improved understanding of knowledge distillation. We employ a control-driven experimental protocol with hypothesis testing and random control distillation to isolate and understand knowledge transfer mechanisms across data modalities. To test the breadth and limits of our analyses, we study self-distillation, standard distillation, feature-map matching variants, distillation scaling laws across model sizes, and the impact of temperature on knowledge transfer. We find statistically supported knowledge transfer in some modalities and architectures; however, the extent of this transfer is less pronounced than anticipated, even under conditions that maximise knowledge sharing. Notably, in cases of significant functional transfer, we identify a consistent and severe asymmetric transfer of negative knowledge to the student, raising safety concerns for knowledge distillation. Across 22 experimental setups, 9 architectures, and 7 datasets, our results suggest that knowledge distillation functions less as a robust compression-by-transfer mechanism and more as a data-dependent regulariser whose transfer component is biased towards negative asymmetric transfer.

📄 PDF Abstract BibTeX arXiv:2510.12615

Code (0)

등록된 구현이 없습니다.

Tasks

Knowledge Distillation

Similar Papers 제목 키워드 기반

What Should Feature Distillation Transfer in LLMs? A Task-Tangent Geometry View

2025-07-14 · Khouloud Saadi, Di Wang arxiv

Feature-based knowledge distillation aims to transfer intermediate representations from a teacher LLM model to a student. Existing approaches typically rely on direct feature matching or learned projections, implicitly t…

Knowledge Distillation

Lipschitz Continuity Guided Knowledge Distillation

2021-08-29 · ICCV 2021 10 · Yuzhang Shang, Bin Duan, Ziliang Zong, Liqiang Nie 외

Knowledge distillation has become one of the most important model compression techniques by distilling knowledge from larger teacher networks to smaller student ones. Although great success has been achieved by prior dis…

Knowledge DistillationModel Compressionobject-detectionObject Detection

Exploring Feature-based Knowledge Distillation for Recommender System: A Frequency Perspective

2024-11-16 · Zhangchi Zhu, Wei zhang

In this paper, we analyze the feature-based knowledge distillation for recommendation from the frequency perspective. By defining knowledge as different frequency components of the features, we theoretically demonstrate …

Knowledge DistillationRecommendation Systems

Multi-perspective Contrastive Logit Distillation

2024-11-16 · Qi Wang, Jinjia Zhou

We propose a novel and efficient logit distillation method, Multi-perspective Contrastive Logit Distillation (MCLD), which leverages contrastive learning to distill logits from multiple perspectives in knowledge distilla…

Contrastive Learningimage-classificationImage ClassificationKnowledge Distillation

On the Orthogonality of Knowledge Distillation with Other Techniques: From an Ensemble Perspective

2020-09-09 · SeongUk Park, KiYoon Yoo, Nojun Kwak

To put a state-of-the-art neural network to practical use, it is necessary to design a model that has a good trade-off between the resource consumption and performance on the test set. Many researchers and engineers are …

Data AugmentationEfficient Neural NetworkKnowledge DistillationQuantization