Saved in:
Bibliographic Details
Main Authors: Feng, Ling, Wu, Tianhao, Ren, Xiangrong, Jing, Zhi, Duan, Xuliang
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2311.13811
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912287540903936
author Feng, Ling
Wu, Tianhao
Ren, Xiangrong
Jing, Zhi
Duan, Xuliang
author_facet Feng, Ling
Wu, Tianhao
Ren, Xiangrong
Jing, Zhi
Duan, Xuliang
contents This paper introduces a new knowledge distillation method, called education distillation (ED), which is inspired by the structured and progressive nature of human learning. ED mimics the educational stages of primary school, middle school, and university and designs teaching reference blocks. The student model is split into a main body and multiple teaching reference blocks to learn from teachers step by step. This promotes efficient knowledge distillation while maintaining the architecture of the student model. Experimental results on the CIFAR100, Tiny Imagenet, Caltech and Food-101 datasets show that the teaching reference blocks can effectively avoid the problem of forgetting. Compared with conventional single-teacher and multi-teacher knowledge distillation methods, ED significantly improves the accuracy and generalization ability of the student model. These findings highlight the potential of ED to improve model performance across different architectures and datasets, indicating its value in various deep learning scenarios. Code examples can be obtained at: https://github.com/Revolutioner1/ED.git.
format Preprint
id arxiv_https___arxiv_org_abs_2311_13811
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Education distillation:getting student models to learn in shcools
Feng, Ling
Wu, Tianhao
Ren, Xiangrong
Jing, Zhi
Duan, Xuliang
Artificial Intelligence
This paper introduces a new knowledge distillation method, called education distillation (ED), which is inspired by the structured and progressive nature of human learning. ED mimics the educational stages of primary school, middle school, and university and designs teaching reference blocks. The student model is split into a main body and multiple teaching reference blocks to learn from teachers step by step. This promotes efficient knowledge distillation while maintaining the architecture of the student model. Experimental results on the CIFAR100, Tiny Imagenet, Caltech and Food-101 datasets show that the teaching reference blocks can effectively avoid the problem of forgetting. Compared with conventional single-teacher and multi-teacher knowledge distillation methods, ED significantly improves the accuracy and generalization ability of the student model. These findings highlight the potential of ED to improve model performance across different architectures and datasets, indicating its value in various deep learning scenarios. Code examples can be obtained at: https://github.com/Revolutioner1/ED.git.
title Education distillation:getting student models to learn in shcools
topic Artificial Intelligence
url https://arxiv.org/abs/2311.13811