EnnengYang / AdaTask

AdaTask: A Task-Aware Adaptive Learning Rate Approach to Multi-Task Learning. AAAI, 2023.

Home Page:https://arxiv.org/abs/2211.15055

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

AdaTask

AdaTask: A Task-Aware Adaptive Learning Rate Approach to Multi-Task Learning. AAAI, 2023.

Abstract

Multi-task learning (MTL) models have demonstrated impressive results in computer vision, natural language processing, and recommender systems. Even though many approaches have been proposed, how well these approaches balance different tasks on each parameter still remains unclear. In this paper, we propose to measure the task dominance degree of a parameter by the total updates of each task on this parameter. Specifically, we compute the total updates by the exponentially decaying Average of the squared Updates (AU) on a parameter from the corresponding task. Based on this novel metric, we observe that many parameters in existing MTL methods, especially those in the higher shared layers, are still dominated by one or several tasks. The dominance of AU is mainly due to the dominance of accumulative gradients from one or several tasks. Motivated by this, we propose a Task-wise Adaptive learning rate approach, AdaTask in short, to separate the accumulative gradients and hence the learning rate of each task for each parameter in adaptive learning rate approaches (e.g., AdaGrad, RMSProp, and Adam). Comprehensive experiments on computer vision and recommender system MTL datasets demonstrate that AdaTask significantly improves the performance of dominated tasks, resulting SOTA average task-wise performance. Analysis on both synthetic and real-world datasets shows AdaTask balance parameters in every shared layer well.

Citation

If you find our paper or this resource helpful, please consider cite:

@article{AdaTask_AAAI2023,
  title={AdaTask: A Task-aware Adaptive Learning Rate Approach to Multi-task Learning},
  author={{Yang, Enneng and Pan, Junwei and Wang, Ximei and Yu, Haibin and Shen, Li and Chen, Xihua and Xiao, Lei and Jiang, Jie and Guo, Guibing},
  journal={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={37},
  number={9},
  pages={10745-10753},
  year={2023}
}

DataSet

  • Download CityScapes dataset and put it in the dataset directory.

Train and Evaluate Method

  python3  main_cityscapes.py --method=adam
  python3  main_cityscapes.py --method=adam_with_adatask

About

AdaTask: A Task-Aware Adaptive Learning Rate Approach to Multi-Task Learning. AAAI, 2023.

https://arxiv.org/abs/2211.15055


Languages

Language:Python 100.0%