paper-with-me

홈 › Papers

What Makes a "Good" Data Augmentation in Knowledge Distillation -- A Statistical Perspective

2020-12-05 · Huan Wang, Suhas Lohit, Mike Jones, Yun Fu

Knowledge distillation (KD) is a general neural network training approach that uses a teacher model to guide the student model. Existing works mainly study KD from the network output side (e.g., trying to design a better KD loss function), while few have attempted to understand it from the input side. Especially, its interplay with data augmentation (DA) has not been well understood. In this paper, we ask: Why do some DA schemes (e.g., CutMix) inherently perform much better than others in KD? What makes a "good" DA in KD? Our investigation from a statistical perspective suggests that a good DA scheme should reduce the covariance of the teacher-student cross-entropy. A practical metric, the stddev of teacher's mean probability (T. stddev), is further presented and well justified empirically. Besides the theoretical understanding, we also introduce a new entropy-based data-mixing DA scheme, CutMixPick, to further enhance CutMix. Extensive empirical studies support our claims and demonstrate how we can harvest considerable performance gains simply by using a better DA scheme in knowledge distillation.

📄 PDF Abstract BibTeX arXiv:2012.02909

Code (1)

mingsun-tse/good-da-in-kd 공식 구현 pytorch

Tasks

Active LearningData AugmentationKnowledge Distillation

Similar Papers 제목 키워드 기반

What Makes Better Augmentation Strategies? Augment Difficult but Not too Different

2021-09-29 · ICLR 2022 4 · Jaehyung Kim, Dongyeop Kang, Sungsoo Ahn, Jinwoo Shin

The practice of data augmentation has been extensively used to boost the performance of deep neural networks for various NLP tasks. It is more effective when only a limited number of labeled samples is available, e.g., l…

Data AugmentationSemantic SimilaritySemantic Textual Similaritytext-classification+1

What Makes a Good Dataset for Knowledge Distillation?

2024-11-19 · CVPR 2025 1 · Logan Frank, Jim Davis

Knowledge distillation (KD) has been a popular and effective method for model compression. One important assumption of KD is that the teacher's original dataset will also be available when training the student. However, …

Continual LearningKnowledge DistillationModel Compression

What makes a good data augmentation for few-shot unsupervised image anomaly detection?

2023-04-06 · Lingrui Zhang, Shuheng Zhang, Guoyang Xie, Jiaqi Liu 외

Data augmentation is a promising technique for unsupervised anomaly detection in industrial applications, where the availability of positive samples is often limited due to factors such as commercial competition and samp…

Anomaly DetectionData AugmentationUnsupervised Anomaly Detection

What makes instance discrimination good for transfer learning?

2020-06-11 · ICLR 2021 1 · Nanxuan Zhao, Zhirong Wu, Rynson W. H. Lau, Stephen Lin

Contrastive visual pretraining based on the instance discrimination pretext task has made significant progress. Notably, recent work on unsupervised pretraining has shown to surpass the supervised counterpart for finetun…

object-detectionObject DetectionTransfer Learning

Teaching What You Should Teach: A Data-Based Distillation Method

2022-12-11 · Shitong Shao, Huanran Chen, Zhen Huang, Linrui Gong 외

In real teaching scenarios, an excellent teacher always teaches what he (or she) is good at but the student is not. This gives the student the best assistance in making up for his (or her) weaknesses and becoming a good …

Data AugmentationKnowledge DistillationObject Recognition