arXiv 2407.140292024-07-19cs.CVcs.LG

PASS++：一种用于无样本类增量学习的双偏差减少框架

PASS++: A Dual Bias Reduction Framework for Non-Exemplar Class-Incremental Learning

Centre for Artificial Intelligence and Robotics, Hong Kong Institute of Science and Innovation, Chinese Academy of Sciences（中国科学院香港创新研究院人工智能与机器人中心）；Institute of Automation of Chinese Academy of Sciences（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）

Fei Zhu, Xu-Yao Zhang, Zhen Cheng, Cheng-Lin Liu

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出PASS++双偏差减少框架，通过自监督变换和原型增强解决无样本类增量学习中的表示与分类器偏差，无需旧样本即可媲美基于样本的方法。

摘要：

类增量学习（CIL）旨在逐步识别新类别，同时保持旧类别的可区分性。大多数现有的CIL方法基于样本存储，即存储部分旧数据用于再训练。在不重新学习旧数据的情况下，这些方法会遭受灾难性遗忘。在本文中，我们发现了CIL中导致旧知识灾难性遗忘的两个固有问题，即表示偏差和分类器偏差。为解决这两个偏差，我们提出了一种简单而新颖的双偏差减少框架，该框架在输入空间采用自监督变换（SST），在深度特征空间采用原型增强（protoAug）。一方面，SST通过学习可跨不同任务迁移的通用且多样的表示来减轻表示偏差。另一方面，protoAug通过在深度特征空间中显式或隐式增强旧类别的原型来克服分类器偏差，从而施加更严格的约束以维持先前学习的决策边界。我们进一步提出了硬度感知的原型增强和多视图集成策略，带来了显著的改进。所提出的框架可以轻松地与预训练模型集成。在不存储任何旧类别样本的情况下，我们的方法可以与存储大量旧数据的最先进的基于样本的方法相媲美。我们希望通过对CIL中存储旧样本的必要性进行重新思考，引起研究人员对无样本CIL的关注。

英文摘要：

Class-incremental learning (CIL) aims to recognize new classes incrementally while maintaining the discriminability of old classes. Most existing CIL methods are exemplar-based, i.e., storing a part of old data for retraining. Without relearning old data, those methods suffer from catastrophic forgetting. In this paper, we figure out two inherent problems in CIL, i.e., representation bias and classifier bias, that cause catastrophic forgetting of old knowledge. To address these two biases, we present a simple and novel dual bias reduction framework that employs self-supervised transformation (SST) in input space and prototype augmentation (protoAug) in deep feature space. On the one hand, SST alleviates the representation bias by learning generic and diverse representations that can transfer across different tasks. On the other hand, protoAug overcomes the classifier bias by explicitly or implicitly augmenting prototypes of old classes in the deep feature space, which poses tighter constraints to maintain previously learned decision boundaries. We further propose hardness-aware prototype augmentation and multi-view ensemble strategies, leading to significant improvements. The proposed framework can be easily integrated with pre-trained models. Without storing any samples of old classes, our method can perform comparably with state-of-the-art exemplar-based approaches which store plenty of old data. We hope to draw the attention of researchers back to non-exemplar CIL by rethinking the necessity of storing old samples in CIL.

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。