arXiv 2405.091332024-05-15cs.LG

克服在线持续学习中的域漂移

Overcoming Domain Drift in Online Continual Learning

Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；Suzhou University of Science and Technology（苏州科技大学）；Youtu Lab, Tencent（腾讯优图实验室）；College of Intelligence and Computing, Tianjin University（天津大学智能与计算学部）

Fan Lyu, Daofeng Liu, Linglan Zhao, Zhang Zhang, Fanhua Shang, Fuyuan Hu, Wei Feng, Liang Wang

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对在线持续学习中存在的灾难性遗忘及持续域漂移导致知识易遗忘的问题，提出漂移减少回放策略，通过代表性样本选择、两级角度跨任务对比边际损失及可选质心蒸馏损失锚定旧任务域，在四个基准数据集上达到SOTA性能。

摘要：

在线持续学习（Online Continual Learning, OCL）使机器学习模型能够在一系列任务中在线获取新知识。然而，OCL面临一个重大挑战：灾难性遗忘，即模型在遇到新任务时，先前任务中学到的知识会被大幅覆盖，导致对先验知识的有偏遗忘。此外，序列学习任务中持续存在的域漂移可能导致已学习特征空间中的决策边界逐渐偏移，使得学到的知识容易被遗忘。为解决上述问题，本文提出一种名为漂移减少回放（Drift-Reducing Rehearsal, DRR）的新型回放策略，用于锚定旧任务的域并减少负迁移效应。首先，我们提出在数据流中通过构造的质心引导下，为更具代表性的样本选择记忆。然后，为防止模型在漂移中陷入域混乱，提出了一种两级角度跨任务对比边际损失（Contrastive Margin Loss, CML），以增强类内和任务内的紧凑性，并增大类间和任务间的差异。最后，为进一步抑制持续域漂移，我们在回放记忆上提出了一种可选的质心蒸馏损失（Centroid Distillation Loss, CDL），以在特征空间中锚定每个先前旧任务的知识。在四个基准数据集上的大量实验结果验证了所提出的DRR能够有效缓解持续域漂移，并在OCL中达到当前最优（state-of-the-art, SOTA）性能。

英文摘要：

Online Continual Learning (OCL) empowers machine learning models to acquire new knowledge online across a sequence of tasks. However, OCL faces a significant challenge: catastrophic forgetting, wherein the model learned in previous tasks is substantially overwritten upon encountering new tasks, leading to a biased forgetting of prior knowledge. Moreover, the continual doman drift in sequential learning tasks may entail the gradual displacement of the decision boundaries in the learned feature space, rendering the learned knowledge susceptible to forgetting. To address the above problem, in this paper, we propose a novel rehearsal strategy, termed Drift-Reducing Rehearsal (DRR), to anchor the domain of old tasks and reduce the negative transfer effects. First, we propose to select memory for more representative samples guided by constructed centroids in a data stream. Then, to keep the model from domain chaos in drifting, a two-level angular cross-task Contrastive Margin Loss (CML) is proposed, to encourage the intra-class and intra-task compactness, and increase the inter-class and inter-task discrepancy. Finally, to further suppress the continual domain drift, we present an optional Centorid Distillation Loss (CDL) on the rehearsal memory to anchor the knowledge in feature space for each previous old task. Extensive experimental results on four benchmark datasets validate that the proposed DRR can effectively mitigate the continual domain drift and achieve the state-of-the-art (SOTA) performance in OCL.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。