Summary of Over-the-air Ensemble Inference with Model Privacy, by Selim F. Yilmaz et al.
Over-the-Air Ensemble Inference with Model Privacy
by Selim F. Yilmaz, Burak Hasircioglu, Deniz Gunduz
First submitted to arxiv on: 7 Feb 2022
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR); Signal Processing (eess.SP)
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 research paper explores distributed inference at the wireless edge, where multiple clients with independent models are queried simultaneously to make an accurate decision on a new sample. The goal is not only to maximize inference accuracy but also to ensure the privacy of local models. To achieve this, the authors propose bandwidth-efficient ensemble inference methods that exploit the superposition property of airwaves. They introduce various over-the-air ensemble methods and demonstrate significant performance improvements compared to orthogonal counterparts while reducing resource usage and providing privacy guarantees. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you have a bunch of friends who each has their own way of solving a math problem. When someone new comes along, they all work together to solve it as accurately as possible, while keeping their own methods private. This is what the researchers did in this study. They developed ways for these “friends” (or clients) to share their knowledge without revealing how they got there. The results show that this approach works better than others and uses fewer resources. |
Keywords
* Artificial intelligence * Inference
