0% found this document useful (0 votes)
7 views2 pages

Knowledge Distillation Techniques Overview

The document provides a comprehensive overview of Knowledge Distillation (KD) techniques, categorizing important research papers into offline, online, deep mutual learning, self-distillation, and ensemble methods. It includes key papers with summaries and links for further reading. Additionally, it features surveys and curated collections related to KD.

Uploaded by

saurav230111002
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
7 views2 pages

Knowledge Distillation Techniques Overview

The document provides a comprehensive overview of Knowledge Distillation (KD) techniques, categorizing important research papers into offline, online, deep mutual learning, self-distillation, and ensemble methods. It includes key papers with summaries and links for further reading. Additionally, it features surveys and curated collections related to KD.

Uploaded by

saurav230111002
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd

========================================

KNOWLEDGE DISTILLATION (KD) TECHNIQUES


========================================

This file lists important research papers for Knowledge Distillation techniques,
including offline, online, deep mutual learning, self-distillation, and ensemble
methods.

------------------------------------------------
1. OFFLINE / CLASSIC KNOWLEDGE DISTILLATION
------------------------------------------------

1. Distilling the Knowledge in a Neural Network


Authors: Geoffrey Hinton, Oriol Vinyals, Jeff Dean
Year: 2015
Summary: Introduced teacher-student knowledge distillation using soft
probabilities.
Link: [Link]

2. Knowledge Distillation: A Survey


Authors: Jianping Gou et al.
Year: 2020
Summary: Comprehensive survey on KD, covering offline methods.
Link: [Link]

------------------------------------------------
2. ONLINE KNOWLEDGE DISTILLATION
------------------------------------------------

3. Knowledge Distillation by On-the-Fly Native Ensemble (ONE)


Authors: Xu Lan, Xiatian Zhu, Shaogang Gong
Year: 2018
Summary: Online KD method with on-the-fly teacher ensemble.
Link: [Link]

4. Peer Collaborative Learning for Online KD


Authors: Guile Wu & Shaogang Gong
Year: 2020
Summary: Peer-based training and ensembling for online KD.
Link: [Link]

5. Decoupled Knowledge with Ensemble Learning for Online Distillation


Authors: Baitan Shao & Ying Chen
Year: 2023
Summary: Improved online distillation using ensemble predictions.
Link: [Link]

------------------------------------------------
3. DEEP MUTUAL LEARNING (DML)
------------------------------------------------

6. Deep Mutual Learning


Authors: Ying Zhang, Tao Xiang, Timothy Hospedales, Huchuan Lu
Year: 2018
Summary: Peer-based online distillation where multiple models learn from each
other.
Link: [Link]

------------------------------------------------
4. SELF-DISTILLATION
------------------------------------------------

7. Be Your Own Teacher: Self Distillation


Authors: Linfeng Zhang et al.
Year: 2019
Summary: Distilling knowledge within the same model.
Link: [Link]

8. Towards Understanding Ensemble, KD & Self-Distillation


Authors: Zeyuan Allen-Zhu & Yuanzhi Li
Year: 2020
Summary: Theoretical analysis connecting KD, self-distillation, and ensemble
effects.
Link: [Link]

------------------------------------------------
5. ENSEMBLE DISTILLATION
------------------------------------------------

9. Online Ensemble Model Compression Using KD


Authors: Devesh Walawalkar et al.
Year: 2020
Summary: Uses ensemble predictions to improve student models.
Link: [Link]

10. Learn from the Past: Experience Ensemble KD


Authors: Chaofei Wang et al.
Year: 2022
Summary: Ensembles intermediate models to improve distillation.
Link: [Link]

------------------------------------------------
6. SURVEYS & PAPER COLLECTIONS
------------------------------------------------

11. Knowledge Distillation: A Survey


Authors: Jianping Gou et al.
Year: 2020
Summary: Comprehensive survey of KD methods including DML, online, self-
distillation, feature-based, and ensemble KD.
Link: [Link]

12. Awesome Knowledge Distillation Paper Collection (GitHub)


Maintainer: Defang Chen
Summary: Curated list of KD papers, including DML, online KD, self-
distillation, adversarial distillation, and feature-based KD.
Link: [Link]

------------------------------------------------
END OF FILE
------------------------------------------------

You might also like