Summary of Heterogeneous Lora For Federated Fine-tuning Of On-device Foundation Models, by Yae Jee Cho and Luyang Liu and Zheng Xu and Aldi Fahrezi and Gauri Joshi
Heterogeneous LoRA for Federated Fine-tuning of On-Device Foundation Models
by Yae Jee Cho, Luyang Liu, Zheng Xu, Aldi Fahrezi, Gauri Joshi
First submitted to arxiv on: 12 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Distributed, Parallel, and Cluster Computing (cs.DC)
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 The proposed HetLoRA method tackles the data and system heterogeneity problem of federated fine-tuning of on-device FMs by using heterogeneous low-rank approximations. This approach allows for efficient aggregation and distribution of heterogeneous LoRA modules, achieving improved convergence speed and final performance compared to homogeneous LoRA. The technique also offers enhanced computation efficiency, making it suitable for federated fine-tuning across heterogeneous devices. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning is a way to train models using data from many different devices without sharing the data itself. On-device FMs are small enough to be deployed on devices for inference, but they can only be fine-tuned with efficient methods. The problem is that devices have different types of data and systems, which makes it hard to fine-tune the model correctly. The HetLoRA method solves this problem by using low-rank approximations in a way that works well across different devices. |
Keywords
* Artificial intelligence * Federated learning * Fine tuning * Inference * Lora
