跳到主要导航 跳到搜索 跳到主要内容

Learning to Generate Gradients for Test-Time Adaptation via Test-Time Training Layers

  • Qi Deng
  • , Shuaicheng Niu
  • , Ronghao Zhang
  • , Yaofo Chen
  • , Runhao Zeng
  • , Jian Chen*
  • , Xiping Hu
  • *此作品的通讯作者
  • South China University of Technology
  • Nanyang Technological University
  • Shenzhen MSU-BIT University

科研成果: 期刊稿件 › 会议文章 › 同行评审

摘要

Test-time adaptation (TTA) aims to fine-tune a trained model online using unlabeled testing data to adapt to new environments or out-of-distribution data, demonstrating broad application potential in real-world scenarios. However, in this optimization process, unsupervised learning objectives like entropy minimization frequently encounter noisy learning signals. These signals produce unreliable gradients, which hinder the model’s ability to converge to an optimal solution quickly and introduce significant instability into the optimization process. In this paper, we seek to resolve these issues from the perspective of optimizer design. Unlike prior TTA using manually designed optimizers like SGD, we employ a learning-to-optimize approach to automatically learn an optimizer, called Meta Gradient Generator (MGG). Specifically, we aim for MGG to effectively utilize historical gradient information during the online optimization process to optimize the current model. To this end, in MGG, we design a lightweight and efficient sequence modeling layer – gradient memory layer. It exploits a self-supervised reconstruction loss to compress historical gradient information into network parameters, thereby enabling better memorization ability over a long-term adaptation process. We only need a small number of unlabeled samples to pre-train MGG, and then the trained MGG can be deployed to process unseen samples. Promising results on ImageNet-C/R/Sketch/A indicate that our method surpasses current state-of-the-art methods with fewer updates, less data, and significantly shorter adaptation times. Compared with a previous SOTA SAR, we achieve 7.4% accuracy improvement and 4.2× faster adaptation speed on ImageNet-C.

源语言英语
页(从-至)16235-16243
页数9
期刊Proceedings of the AAAI Conference on Artificial Intelligence
卷39
期15
DOI
出版状态已出版 - 11 4月 2025
已对外发布是
活动39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025 - Philadelphia, 美国
期限: 25 2月 2025 → 4 3月 2025

学术指纹

探究 'Learning to Generate Gradients for Test-Time Adaptation via Test-Time Training Layers' 的科研主题。它们共同构成独一无二的学术指纹。

引用此