Saved in:
Bibliographic Details
Main Authors: Xuan, Xiwei, Deng, Ziquan, Lin, Hsuan-Tien, Ma, Kwan-Liu
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.05594
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913420726501376
author Xuan, Xiwei
Deng, Ziquan
Lin, Hsuan-Tien
Ma, Kwan-Liu
author_facet Xuan, Xiwei
Deng, Ziquan
Lin, Hsuan-Tien
Ma, Kwan-Liu
contents Recent studies highlight that deep learning models often learn spurious features mistakenly linked to labels, compromising their reliability in real-world scenarios where such correlations do not hold. Despite the increasing research effort, existing solutions often face two main challenges: they either demand substantial annotations of spurious attributes, or they yield less competitive outcomes with expensive training when additional annotations are absent. In this paper, we introduce SLIM, a cost-effective and performance-targeted approach to reducing spurious correlations in deep learning. Our method leverages a human-in-the-loop protocol featuring a novel attention labeling mechanism with a constructed attention representation space. SLIM significantly reduces the need for exhaustive additional labeling, requiring human input for fewer than 3% of instances. By prioritizing data quality over complicated training strategies, SLIM curates a smaller yet more feature-balanced data subset, fostering the development of spuriousness-robust models. Experimental validations across key benchmarks demonstrate that SLIM competes with or exceeds the performance of leading methods while significantly reducing costs. The SLIM framework thus presents a promising path for developing reliable models more efficiently. Our code is available in https://github.com/xiweix/SLIM.git/.
format Preprint
id arxiv_https___arxiv_org_abs_2407_05594
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SLIM: Spuriousness Mitigation with Minimal Human Annotations
Xuan, Xiwei
Deng, Ziquan
Lin, Hsuan-Tien
Ma, Kwan-Liu
Computer Vision and Pattern Recognition
Recent studies highlight that deep learning models often learn spurious features mistakenly linked to labels, compromising their reliability in real-world scenarios where such correlations do not hold. Despite the increasing research effort, existing solutions often face two main challenges: they either demand substantial annotations of spurious attributes, or they yield less competitive outcomes with expensive training when additional annotations are absent. In this paper, we introduce SLIM, a cost-effective and performance-targeted approach to reducing spurious correlations in deep learning. Our method leverages a human-in-the-loop protocol featuring a novel attention labeling mechanism with a constructed attention representation space. SLIM significantly reduces the need for exhaustive additional labeling, requiring human input for fewer than 3% of instances. By prioritizing data quality over complicated training strategies, SLIM curates a smaller yet more feature-balanced data subset, fostering the development of spuriousness-robust models. Experimental validations across key benchmarks demonstrate that SLIM competes with or exceeds the performance of leading methods while significantly reducing costs. The SLIM framework thus presents a promising path for developing reliable models more efficiently. Our code is available in https://github.com/xiweix/SLIM.git/.
title SLIM: Spuriousness Mitigation with Minimal Human Annotations
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.05594