Class-Incremental Learning: A Survey

Abstract

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. 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 class-incremental learning and summarize these methods from several aspects. We also provide a rigorous and unified evaluation of 17 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 is available at https://github.com/zhoudw-zdw/CIL_Survey/

Accepted to TPAMI. Code is available at https://github.com/zhoudw-zdw/CIL_Survey/

References in corpus (10)

• Distilling the Knowledge in a Neural Network
• Conditional Generative Adversarial Nets
• Learning under Concept Drift: A Review
• A Comprehensive Study of Class Incremental Learning Algorithms for Visual Tasks
• Continual Lifelong Learning in Natural Language Processing: A Survey
• Uncertainty-aware Contrastive Distillation for Incremental Semantic Segmentation
• Online Hyperparameter Optimization for Class-Incremental Learning
• Continual Diffusion: Continual Customization of Text-to-Image Diffusion with C-LoRA
• Exploring Continual Learning of Diffusion Models
• Continual Learning by Modeling Intra-Class Variation

Cited by in corpus (9)

• A Survey on Continual Semantic Segmentation: Theory, Challenge, Method and Application
• Learning without Forgetting for Vision-Language Models
• Advances in Continual Graph Learning for Anti-Money Laundering Systems: A Comprehensive Review
• Continual Learning with Neuromorphic Computing: Foundations, Methods, and Emerging Applications
• Adaptive Intrusion Detection for Evolving RPL IoT Attacks Using Incremental Learning
• Efficient and Robust Continual Graph Learning for Graph Classification in Biology
• Buffer-free Class-Incremental Learning with Out-of-Distribution Detection
• Improving Forward Compatibility in Class Incremental Learning by Increasing Representation Rank and Feature Richness
• Developing a Strong CPS Defender: An Evolutionary Approach