LFME: A Simple Framework for Learning from Multiple Experts in Domain Generalization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Liang, Zhang, Yong, Song, Yibing, Shen, Zhiqiang, Liu, Lingqiao
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929557902196736
author Chen, Liang
Zhang, Yong
Song, Yibing
Shen, Zhiqiang
Liu, Lingqiao
author_facet Chen, Liang
Zhang, Yong
Song, Yibing
Shen, Zhiqiang
Liu, Lingqiao
contents Domain generalization (DG) methods aim to maintain good performance in an unseen target domain by using training data from multiple source domains. While success on certain occasions are observed, enhancing the baseline across most scenarios remains challenging. This work introduces a simple yet effective framework, dubbed learning from multiple experts (LFME), that aims to make the target model an expert in all source domains to improve DG. Specifically, besides learning the target model used in inference, LFME will also train multiple experts specialized in different domains, whose output probabilities provide professional guidance by simply regularizing the logit of the target model. Delving deep into the framework, we reveal that the introduced logit regularization term implicitly provides effects of enabling the target model to harness more information, and mining hard samples from the experts during training. Extensive experiments on benchmarks from different DG tasks demonstrate that LFME is consistently beneficial to the baseline and can achieve comparable performance to existing arts. Code is available at~\url{https://github.com/liangchen527/LFME}.
format Preprint
id arxiv_https___arxiv_org_abs_2410_17020
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LFME: A Simple Framework for Learning from Multiple Experts in Domain Generalization
Chen, Liang
Zhang, Yong
Song, Yibing
Shen, Zhiqiang
Liu, Lingqiao
Machine Learning
Computer Vision and Pattern Recognition
Domain generalization (DG) methods aim to maintain good performance in an unseen target domain by using training data from multiple source domains. While success on certain occasions are observed, enhancing the baseline across most scenarios remains challenging. This work introduces a simple yet effective framework, dubbed learning from multiple experts (LFME), that aims to make the target model an expert in all source domains to improve DG. Specifically, besides learning the target model used in inference, LFME will also train multiple experts specialized in different domains, whose output probabilities provide professional guidance by simply regularizing the logit of the target model. Delving deep into the framework, we reveal that the introduced logit regularization term implicitly provides effects of enabling the target model to harness more information, and mining hard samples from the experts during training. Extensive experiments on benchmarks from different DG tasks demonstrate that LFME is consistently beneficial to the baseline and can achieve comparable performance to existing arts. Code is available at~\url{https://github.com/liangchen527/LFME}.
title LFME: A Simple Framework for Learning from Multiple Experts in Domain Generalization
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.17020