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：深层模型,例如CNN和愿景变异器等,在封闭世界的许多愿景任务中取得了令人印象深刻的成就。然而,在我们不断变化的世界中,新课程不时出现,要求不断获得新知识。例如,机器人需要了解新的指示,意见监测系统应每天分析新专题。 级入门学习(CIL)使学习者能够逐步纳入新课程知识,并在所见的所有类别中建立起一个通用的分级器。相应的是,当直接以新的类别实例培训模型时,出现了一个致命的问题 -- -- 模型往往灾难性地忘记了以前模式的特点,其性能也急剧下降。例如,一个机器人需要了解新的指示,而一个意见监测系统则需要每天分析新出现的主题。 类入门学习(CIL)使学习者能够从三个方面,即数据中心、模型中心、算法中心、以及算法中心等中总结出这些方法。 我们还对基准图像分类任务的16种方法进行了严格和统一的评价,以找出不同算法的特征,而模型的特征及其性能急剧下降。在机器学习界中,我们已作出许多努力,从而忽略了当前预算上的比较结果。