Summary of Federated Aggregation Of Mallows Rankings: a Comparative Analysis Of Borda and Lehmer Coding, by Jin Sima et al.
Federated Aggregation of Mallows Rankings: A Comparative Analysis of Borda and Lehmer Coding
by Jin Sima, Vishal Rana, Olgica Milenkovic
First submitted to arxiv on: 1 Sep 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 This paper presents the first federated rank aggregation protocols for distributed learning across multiple clients. The proposed methods, Borda scoring and Lehmer coding, enable private and communication-efficient learning in fields like biomedical data sharing, where rankings are distributed and require privacy. The authors analyze the sample complexity of these methods on Mallows distributions with a known scaling factor and unknown centroid permutation. They show that for certain conditions, each client needs to locally aggregate a fixed number of rankings, quantize the result, and send it to the server to recover the centroid permutation with high probability. The communication complexity scales linearly with the number of clients and data points. These protocols have implications for distributed machine learning applications in various domains. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to combine rankings from different sources so that they can be kept private. Imagine you’re working on a big project with many people, and each person has their own list of the best solutions. You want to combine these lists into one master list, but you don’t want to reveal your personal ranking. This paper presents two methods to do this: Borda scoring and Lehmer coding. These methods allow you to combine rankings without revealing too much about individual people’s preferences. The authors show that their methods work well in certain situations, which is important for many applications, like sharing data in the biomedical field. |
Keywords
» Artificial intelligence » Machine learning » Probability
