Summary of Fedkbp: Federated Dose Prediction Framework For Knowledge-based Planning in Radiation Therapy, by Jingyun Chen et al.
FedKBP: Federated dose prediction framework for knowledge-based planning in radiation therapy
by Jingyun Chen, Martin King, Yading Yuan
First submitted to arxiv on: 17 Aug 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 recent breakthrough in deep learning-based dose prediction methods has necessitated collaboration among medical centers to jointly train models without compromising patient data privacy. The FedKBP framework was developed to evaluate the performances of centralized, federated, and individual training of dose prediction models on the OpenKBP dataset. To simulate FL and individual training, the data was divided into 8 training sites with two types of case distributions: IID and non-IID. The results show FL consistently outperforms individual training in terms of model optimization speed and out-of-sample testing scores. Under IID data division, FL shows comparable performance to centralized training, while under non-IID division, larger sites outperform smaller sites by up to 19%. However, non-IID FL shows reduced performance compared to IID FL, highlighting the need for more sophisticated FL methods beyond mere model averaging. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning is a new way for medical centers to work together without sharing patient data. It helps improve the accuracy of dose prediction models by combining information from different sites. The study looked at how well this approach works compared to other ways of training models, such as doing it all in one place or letting each site do its own thing. They found that working together (federated learning) is better than any of these other approaches. It’s like a team effort that helps get the best results. |
Keywords
» Artificial intelligence » Deep learning » Federated learning » Optimization
