Summary of Harmonizing Generalization and Personalization in Federated Prompt Learning, by Tianyu Cui et al.
Harmonizing Generalization and Personalization in Federated Prompt Learning
by Tianyu Cui, Hongxia Li, Jingya Wang, Ye Shi
First submitted to arxiv on: 16 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 This paper proposes Federated Prompt Learning (FPL) that incorporates large pre-trained Vision-Language models (VLMs) into federated learning through prompt tuning. By leveraging the transferable representations and remarkable generalization capacity of VLMs, FPL aims to strike a balance between personalization and generalization in addressing data heterogeneity. The proposed approach, called Federated Prompt Learning with CLIP Generalization and low-rank Personalization (FedPGP), employs pre-trained CLIP to provide knowledge-guidance on the global prompt for improved generalization and incorporates a low-rank adaptation term to personalize the global prompt. FedPGP also integrates a prompt-wise contrastive loss to achieve knowledge guidance and personalized adaptation simultaneously, enabling a harmonious balance between personalization and generalization in FPL. Experimental results on various datasets demonstrate the superiority of FedPGP in balancing generalization and personalization. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper talks about how to make machines learn from different sources of information. It’s like trying to get different people to agree on something! The researchers came up with a new way to do this, called Federated Prompt Learning (FPL). FPL uses big pre-trained models that can understand pictures and text, and combines them with a special kind of learning called federated learning. This helps machines learn from all the different sources of information without getting confused. The researchers tested their idea on many different datasets and showed that it works really well. |
Keywords
» Artificial intelligence » Contrastive loss » Federated learning » Generalization » Low rank adaptation » Prompt
