paper-with-me

Papers

Does Knowledge Distillation Really Work?

2021-06-10 · NeurIPS 2021 12 · Samuel Stanton, Pavel Izmailov, Polina Kirichenko, Alexander A. Alemi, Andrew Gordon Wilson

Knowledge distillation is a popular technique for training a small student network to emulate a larger teacher model, such as an ensemble of networks. We show that while knowledge distillation can improve student generalization, it does not typically work as it is commonly understood: there often remains a surprisingly large discrepancy between the predictive distributions of the teacher and the student, even in cases when the student has the capacity to perfectly match the teacher. We identify difficulties in optimization as a key reason for why the student is unable to match the teacher. We also show how the details of the dataset used for distillation play a role in how closely the student matches the teacher -- and that more closely matching the teacher paradoxically does not always lead to better student generalization.

📄 PDF Abstract BibTeX arXiv:2106.05945

Code (2)

samuelstanton/gnosis 공식 구현 pytorch
jincan333/lot pytorch

Tasks

Knowledge Distillation

Methods 이 논문이 사용한 방법론

Knowledge Distillation A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions.…

Similar Papers 제목 키워드 기반

Why does Knowledge Distillation Work? Rethink its Attention and Fidelity Mechanism

2024-04-30 · Chenqi Guo, Shiwei Zhong, Xiaofeng Liu, Qianli Feng 외

Does Knowledge Distillation (KD) really work? Conventional wisdom viewed it as a knowledge transfer procedure where a perfect mimicry of the student to its teacher is desired. However, paradoxical studies indicate that c…

Data AugmentationDiversityKnowledge DistillationTransfer Learning

Tree-Like Decision Distillation

2021-06-19 · CVPR 2021 1 · Jie Song, Haofei Zhang, Xinchao Wang, Mengqi Xue 외

Knowledge distillation pursues a diminutive yet well-behaved student network by harnessing the knowledge learned by a cumbersome teacher model. Prior methods achieve this by making the student imitate shallow behavio…

Decision MakingKnowledge Distillation

CleverDistiller: Simple and Spatially Consistent Cross-modal Distillation

2025-03-12 · Hariprasath Govindarajan, Maciej K. Wozniak, Marvin Klingner, Camille Maurice 외

Vision foundation models (VFMs) such as DINO have led to a paradigm shift in 2D camera-based perception towards extracting generalized features to support many downstream tasks. Recent works introduce self-supervised cro…

3D Object DetectionAutonomous DrivingKnowledge Distillationobject-detection+4

Anti-Distillation Backdoor Attacks: Backdoors Can Really Survive in Knowledge Distillation

2021-10-24 · MM - Proceedings of the ACM International Conference on Multimedia 2021 10 · Yunjie Ge, Qian Wang, Baolin Zheng, Xinlu Zhuang 외

Motivated by resource-limited scenarios, knowledge distillation (KD) has received growing attention, effectively and quickly producing lightweight yet high-performance student models by transferring the dark knowledge fr…

Backdoor AttackKnowledge Distillation

What Knowledge Gets Distilled in Knowledge Distillation?

2022-05-31 · NeurIPS 2023 11

Knowledge distillation aims to transfer useful information from a teacher network to a student network, with the primary goal of improving the student's performance for the task at hand. Over the years, there has a been …

Knowledge Distillation