Summary of Priphit: Privacy-preserving Hierarchical Training Of Deep Neural Networks, by Yamin Sepehri et al.
PriPHiT: Privacy-Preserving Hierarchical Training of Deep Neural Networks
by Yamin Sepehri, Pedram Pad, Pascal Frossard, L. Andrea Dunbar
First submitted to arxiv on: 9 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Cryptography and Security (cs.CR); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
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 method to train deep neural networks on both edge devices and cloud servers while preserving privacy by suppressing sensitive content at the edge device. The approach uses adversarial early exits to transmit task-relevant information to the cloud, incorporating noise addition during training to ensure differential privacy. The authors test their method on various facial and medical datasets with different architectures, demonstrating its performance. Additionally, they showcase successful defenses against reconstruction attacks. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about finding a way to train computers using deep learning without sharing personal or private information. Imagine you’re sending a picture of your face for training, but you don’t want anyone else to see it. This method helps by hiding the sensitive parts and only sending what’s important. It works really well and even protects against people trying to get the hidden information back. |
Keywords
* Artificial intelligence * Deep learning