This repository serves as a comprehensive catalog for incremental learning (IL), also known as continual learning or lifelong learning. It targets researchers and practitioners in machine learning by providing a curated collection of papers, code, datasets, and talks related to mitigating catastrophic forgetting and enabling models to learn sequentially. The primary benefit is a centralized, up-to-date resource for staying abreast of advancements in this rapidly evolving field.

How It Works

The repository functions as a structured index, categorizing resources by topic, year, and approach (e.g., rehearsal, generative replay, parameter isolation). It links to seminal papers, recent conference contributions, and practical toolkits like CLHive, PTIL, FACIL, Avalanche, and PyCIL, offering a broad overview of the IL landscape. The organization facilitates discovery of relevant research and implementation resources.

Quick Start & Requirements

This repository is a curated list of resources, not a runnable software package. It requires no installation. Accessing the linked papers, code repositories, and datasets will have their own respective requirements.

Highlighted Details

• Extensive catalog of papers from major conferences (CVPR, NeurIPS, ICLR, ECCV, ICML) dating back to 2016, organized by year and category.
• Links to popular IL frameworks and toolboxes, including Avalanche, PyCIL, CLHive, and LAMDA-PILOT.
• Curated lists of datasets commonly used in IL research, such as ImageNet, CIFAR-10/100, and OpenLORIS-Object.
• Resources for learning, including tutorials, workshops, and talks from prominent researchers.

Maintenance & Community

The repository is maintained by the Vision-Intelligence-and-Robots-Group. Contact information for concerns is provided (hongxiaopeng@ieee.org, xl330@126.com). The project appears to be actively updated, with recent entries from 2024.

Licensing & Compatibility

The repository itself does not specify a license. Individual linked code repositories and datasets will have their own licenses.

Limitations & Caveats

This is a curated list and does not provide a unified API or framework for implementing incremental learning methods. Users must navigate to individual linked resources for code execution and further details.