https://github.com/amazon-science/lc-loss
Science Score: 23.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
○.zenodo.json file
-
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (11.0%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: amazon-science
- License: other
- Language: Python
- Default Branch: main
- Size: 154 KB
Statistics
- Stars: 2
- Watchers: 3
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
Avoiding spurious correlations via logit correction
This repository provides the official PyTorch implementation of the core algorithms (including logit correction and groupMixUp) in the following paper:
Avoiding spurious correlations via logit correction
Sheng Liu (NYU), Xu Zhang (Amazon), Nitesh Sekhar (Amazon), Yue Wu (Amazon), Prateek Singhal (Amazon), Carlos Fernandez-Granda (NYU) ICLR 2023Paper: Arxiv
Abstract:
Empirical studies suggest that machine learning models trained with empirical risk minimization (ERM) often rely on attributes that may be spuriously correlated with the class labels. Such models typically lead to poor performance during inference for data lacking such correlations. In this work, we explicitly consider a situation where potential spurious correlations are present in the majority of training data. In contrast with existing approaches, which use the ERM model outputs to detect the samples without spurious correlations and either heuristically upweight or upsample those samples, we propose the logit correction (LC) loss, a simple yet effective improvement on the softmax cross-entropy loss, to correct the sample logit. We demonstrate that minimizing the LC loss is equivalent to maximizing the group-balanced accuracy, so the proposed LC could mitigate the negative impacts of spurious correlations. Our extensive experimental results further reveal that the proposed LC loss outperforms state-of-the-art solutions on multiple popular benchmarks by a large margin, an average 5.5% absolute improvement, without access to spurious attribute labels. LC is also competitive with oracle methods that make use of the attribute labels.
For a full E2E training and evaluation example, please refer to link
Citations
Please cite our paper if you find it to be helpful in your research.
BibTeX
bibtex
@Inproceedings{
Liu2023,
author = {Sheng Liu and Xu Zhang and Nitesh Sekhar and Yue Wu and Prateek Singhal and Carlos Fernandez-Granda},
title = {Avoiding spurious correlations via logit correction},
year = {2023},
url = {https://www.amazon.science/publications/avoiding-spurious-correlations-via-logit-correction},
booktitle = {ICLR 2023},
}
Contact
Sheng Liu (shengliu@nyu.edu), Xu Zhang (xzhnamz@amazon.com)
Acknowledgments
This work was mainly done the authors was doing internship at Amazon Science.
Owner
- Name: Amazon Science
- Login: amazon-science
- Kind: organization
- Website: https://amazon.science
- Twitter: AmazonScience
- Repositories: 80
- Profile: https://github.com/amazon-science
GitHub Events
Total
Last Year
Issues and Pull Requests
Last synced: over 1 year ago
All Time
- Total issues: 2
- Total pull requests: 0
- Average time to close issues: about 19 hours
- Average time to close pull requests: N/A
- Total issue authors: 1
- Total pull request authors: 0
- Average comments per issue: 1.0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 0
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 0
- Pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
- YeonsungJung (1)