Summary of Pip: Prototypes-injected Prompt For Federated Class Incremental Learning, by Muhammad Anwar Ma’sum et al.
PIP: Prototypes-Injected Prompt for Federated Class Incremental Learning
by Muhammad Anwar Ma’sum, Mahardhika Pratama, Savitha Ramasamy, Lin Liu, Habibullah Habibullah, Ryszard Kowalczyk
First submitted to arxiv on: 30 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel Federated Class Incremental Learning (FCIL) method, called prototypes-injected prompt (PIP), is proposed to address catastrophic forgetting and non-IID data distribution simultaneously. The PIP method involves prototype injection on prompt learning, prototype augmentation, and weighted Gaussian aggregation on the server side. It outperforms current state-of-the-art methods by up to 33% in CIFAR100, MiniImageNet, and TinyImageNet datasets. The proposed method requires smaller participating local clients and global rounds, making it a robust solution for different task sizes. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary FCIL is a new way to keep learning without forgetting old things. It’s hard because data isn’t always the same everywhere. A team came up with a new idea called PIP that makes FCIL work better. They did some tests and showed that PIP works much better than other methods, especially on pictures. This means computers can learn more from different places and not forget what they learned before. |
Keywords
» Artificial intelligence » Prompt
