Deep models, e.g., CNNs and Vision Transformers, have achieved impressive achievements in many vision tasks in the closed world. However, novel classes emerge from time to time in our ever-changing world, requiring a learning system to acquire new knowledge continually. For example, a robot needs to understand new instructions, and an opinion monitoring system should analyze emerging topics every day. Class-Incremental Learning (CIL) enables the learner to incorporate the knowledge of new classes incrementally and build a universal classifier among all seen classes. Correspondingly, when directly training the model with new class instances, a fatal problem occurs -- the model tends to catastrophically forget the characteristics of former ones, and its performance drastically degrades. There have been numerous efforts to tackle catastrophic forgetting in the machine learning community. In this paper, we survey comprehensively recent advances in deep class-incremental learning and summarize these methods from three aspects, i.e., data-centric, model-centric, and algorithm-centric. We also provide a rigorous and unified evaluation of 16 methods in benchmark image classification tasks to find out the characteristics of different algorithms empirically. Furthermore, we notice that the current comparison protocol ignores the influence of memory budget in model storage, which may result in unfair comparison and biased results. Hence, we advocate fair comparison by aligning the memory budget in evaluation, as well as several memory-agnostic performance measures. The source code to reproduce these evaluations is available at https://github.com/zhoudw-zdw/CIL_Survey/

翻译：深度模型，例如卷积神经网络（CNN）和视觉Transformer，在封闭世界的许多视觉任务中取得了令人瞩目的成就。然而，在我们不断变化的世界中，新类别会不时涌现，这要求学习系统持续获取新知识。例如，机器人需要理解新的指令，舆情监控系统则应每天分析新出现的话题。类增量学习（CIL）使学习者能够逐步融入新类别的知识，并在所有已见类别中构建通用分类器。相应地，当直接用新类别实例训练模型时，会出现一个致命问题——模型倾向于灾难性地遗忘先前类别的特征，导致其性能急剧下降。机器学习界已投入大量努力来解决灾难性遗忘问题。本文全面调研了深度类增量学习的最新进展，并从三个方面对这些方法进行归纳，即数据为中心、模型为中心和算法为中心。我们还对16种方法在基准图像分类任务中进行了严格且统一的评估，以实证地发现不同算法的特性。此外，我们注意到当前的比较协议忽略了内存预算对模型存储的影响，这可能导致不公平的比较和有偏差的结果。因此，我们主张通过在评估中统一内存预算以及采用若干内存无关的性能度量来实现公平比较。重现这些评估的源代码可在 https://github.com/zhoudw-zdw/CIL_Survey/ 获取。