GO4Align: Group Optimization for Multi-Task Alignment
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_ | 1866913566632706048 |
|---|---|
| author | Shen, Jiayi Wang, Cheems Xiao, Zehao Van Noord, Nanne Worring, Marcel |
| author_facet | Shen, Jiayi Wang, Cheems Xiao, Zehao Van Noord, Nanne Worring, Marcel |
| contents | This paper proposes \textit{GO4Align}, a multi-task optimization approach that tackles task imbalance by explicitly aligning the optimization across tasks. To achieve this, we design an adaptive group risk minimization strategy, comprising two techniques in implementation: (i) dynamical group assignment, which clusters similar tasks based on task interactions; (ii) risk-guided group indicators, which exploit consistent task correlations with risk information from previous iterations. Comprehensive experimental results on diverse benchmarks demonstrate our method's performance superiority with even lower computational costs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_06486 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | GO4Align: Group Optimization for Multi-Task Alignment Shen, Jiayi Wang, Cheems Xiao, Zehao Van Noord, Nanne Worring, Marcel Machine Learning Computer Vision and Pattern Recognition This paper proposes \textit{GO4Align}, a multi-task optimization approach that tackles task imbalance by explicitly aligning the optimization across tasks. To achieve this, we design an adaptive group risk minimization strategy, comprising two techniques in implementation: (i) dynamical group assignment, which clusters similar tasks based on task interactions; (ii) risk-guided group indicators, which exploit consistent task correlations with risk information from previous iterations. Comprehensive experimental results on diverse benchmarks demonstrate our method's performance superiority with even lower computational costs. |
| title | GO4Align: Group Optimization for Multi-Task Alignment |
| topic | Machine Learning Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.06486 |