Awesome Incremental Learning is a curated collection of research papers and resources focused on incremental learning and continual learning approaches.
The collection addresses the challenge of how machine learning systems can learn from new data over time without forgetting previously acquired knowledge, a problem known as catastrophic forgetting. The resource organizes research across multiple dimensions of this problem, including class-incremental learning where systems learn new classes sequentially, continual learning with pre-trained models, online learning scenarios, and domain-specific applications in natural language processing, computer vision, and recommender systems. It also covers emerging areas such as continual learning for vision-language models and large language models.
Developers and researchers should use this collection to understand the landscape of incremental learning research and identify relevant papers for their specific use case. The resource is particularly valuable for those working on systems that must adapt to new data without retraining from scratch, whether in computer vision, NLP, or multimodal settings. The breadth of surveys and papers makes it suitable for both newcomers seeking foundational understanding and practitioners looking for state-of-the-art methods in specific incremental learning scenarios.
The project maintains an organized index of peer-reviewed research spanning multiple years and venues, with papers sourced from top-tier conferences and journals. The collection includes links to both paper references and associated code repositories where available, supporting reproducibility and practical implementation of the methods discussed.
