Learning Decomposable and Debiased Representations via Attribute-Centric Information Bottlenecks
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913276278865920 |
|---|---|
| author | Hong, Jinyung Jeon, Eun Som Kim, Changhoon Park, Keun Hee Nath, Utkarsh Yang, Yezhou Turaga, Pavan Pavlic, Theodore P. |
| author_facet | Hong, Jinyung Jeon, Eun Som Kim, Changhoon Park, Keun Hee Nath, Utkarsh Yang, Yezhou Turaga, Pavan Pavlic, Theodore P. |
| contents | Biased attributes, spuriously correlated with target labels in a dataset, can problematically lead to neural networks that learn improper shortcuts for classifications and limit their capabilities for out-of-distribution (OOD) generalization. Although many debiasing approaches have been proposed to ensure correct predictions from biased datasets, few studies have considered learning latent embedding consisting of intrinsic and biased attributes that contribute to improved performance and explain how the model pays attention to attributes. In this paper, we propose a novel debiasing framework, Debiasing Global Workspace, introducing attention-based information bottlenecks for learning compositional representations of attributes without defining specific bias types. Based on our observation that learning shape-centric representation helps robust performance on OOD datasets, we adopt those abilities to learn robust and generalizable representations of decomposable latent embeddings corresponding to intrinsic and biasing attributes. We conduct comprehensive evaluations on biased datasets, along with both quantitative and qualitative analyses, to showcase our approach's efficacy in attribute-centric representation learning and its ability to differentiate between intrinsic and bias-related features. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_14140 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Learning Decomposable and Debiased Representations via Attribute-Centric Information Bottlenecks Hong, Jinyung Jeon, Eun Som Kim, Changhoon Park, Keun Hee Nath, Utkarsh Yang, Yezhou Turaga, Pavan Pavlic, Theodore P. Computer Vision and Pattern Recognition Machine Learning Biased attributes, spuriously correlated with target labels in a dataset, can problematically lead to neural networks that learn improper shortcuts for classifications and limit their capabilities for out-of-distribution (OOD) generalization. Although many debiasing approaches have been proposed to ensure correct predictions from biased datasets, few studies have considered learning latent embedding consisting of intrinsic and biased attributes that contribute to improved performance and explain how the model pays attention to attributes. In this paper, we propose a novel debiasing framework, Debiasing Global Workspace, introducing attention-based information bottlenecks for learning compositional representations of attributes without defining specific bias types. Based on our observation that learning shape-centric representation helps robust performance on OOD datasets, we adopt those abilities to learn robust and generalizable representations of decomposable latent embeddings corresponding to intrinsic and biasing attributes. We conduct comprehensive evaluations on biased datasets, along with both quantitative and qualitative analyses, to showcase our approach's efficacy in attribute-centric representation learning and its ability to differentiate between intrinsic and bias-related features. |
| title | Learning Decomposable and Debiased Representations via Attribute-Centric Information Bottlenecks |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2403.14140 |