Saved in:
Bibliographic Details
Main Authors: Kim, Hyungmin, Suh, Sungho, Baek, Sunghyun, Kim, Daehwan, Jeong, Daun, Cho, Hansang, Kim, Junmo
Format: Preprint
Published: 2022
Subjects:
Online Access:https://arxiv.org/abs/2211.10938
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910376415723520
author Kim, Hyungmin
Suh, Sungho
Baek, Sunghyun
Kim, Daehwan
Jeong, Daun
Cho, Hansang
Kim, Junmo
author_facet Kim, Hyungmin
Suh, Sungho
Baek, Sunghyun
Kim, Daehwan
Jeong, Daun
Cho, Hansang
Kim, Junmo
contents We present a novel adversarial penalized self-knowledge distillation method, named adversarial learning and implicit regularization for self-knowledge distillation (AI-KD), which regularizes the training procedure by adversarial learning and implicit distillations. Our model not only distills the deterministic and progressive knowledge which are from the pre-trained and previous epoch predictive probabilities but also transfers the knowledge of the deterministic predictive distributions using adversarial learning. The motivation is that the self-knowledge distillation methods regularize the predictive probabilities with soft targets, but the exact distributions may be hard to predict. Our method deploys a discriminator to distinguish the distributions between the pre-trained and student models while the student model is trained to fool the discriminator in the trained procedure. Thus, the student model not only can learn the pre-trained model's predictive probabilities but also align the distributions between the pre-trained and student models. We demonstrate the effectiveness of the proposed method with network architectures on multiple datasets and show the proposed method achieves better performance than state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2211_10938
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle AI-KD: Adversarial learning and Implicit regularization for self-Knowledge Distillation
Kim, Hyungmin
Suh, Sungho
Baek, Sunghyun
Kim, Daehwan
Jeong, Daun
Cho, Hansang
Kim, Junmo
Computer Vision and Pattern Recognition
Machine Learning
We present a novel adversarial penalized self-knowledge distillation method, named adversarial learning and implicit regularization for self-knowledge distillation (AI-KD), which regularizes the training procedure by adversarial learning and implicit distillations. Our model not only distills the deterministic and progressive knowledge which are from the pre-trained and previous epoch predictive probabilities but also transfers the knowledge of the deterministic predictive distributions using adversarial learning. The motivation is that the self-knowledge distillation methods regularize the predictive probabilities with soft targets, but the exact distributions may be hard to predict. Our method deploys a discriminator to distinguish the distributions between the pre-trained and student models while the student model is trained to fool the discriminator in the trained procedure. Thus, the student model not only can learn the pre-trained model's predictive probabilities but also align the distributions between the pre-trained and student models. We demonstrate the effectiveness of the proposed method with network architectures on multiple datasets and show the proposed method achieves better performance than state-of-the-art methods.
title AI-KD: Adversarial learning and Implicit regularization for self-Knowledge Distillation
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2211.10938