Summary of Hypernetwork-driven Model Fusion For Federated Domain Generalization, by Marc Bartholet et al.
Hypernetwork-Driven Model Fusion for Federated Domain Generalization
by Marc Bartholet, Taehyeon Kim, Ami Beuret, Se-Young Yun, Joachim M. Buhmann
First submitted to arxiv on: 10 Feb 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 a novel framework, called hypernetwork-based Federated Fusion (hFedF), to address the challenges of domain shifts in heterogeneous data in Federated Learning (FL). The traditional approach of linear model averaging is limited by its inability to learn domain-invariant features. hFedF uses hypernetworks for non-linear aggregation, facilitating generalization to unseen domains. The method employs client-specific embeddings and gradient alignment techniques to manage domain generalization effectively. The results demonstrate superior performance in handling domain shifts in both zero-shot and few-shot settings. Comprehensive comparisons on PACS, Office-Home, and VLCS datasets show that hFedF consistently achieves the highest in-domain and out-of-domain accuracy with reliable predictions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated Learning (FL) helps many devices learn together without sharing their data. But when this data is very different, performance drops. A new way to fix this problem uses hypernetworks to mix up learning from different places. This makes it better at generalizing to new situations. The results show that this method works well in both quick and slow learning settings. It also outperforms other methods on three big datasets. |
Keywords
* Artificial intelligence * Alignment * Domain generalization * Federated learning * Few shot * Generalization * Zero shot
