Summary of Training Differentially Private Ad Prediction Models with Semi-sensitive Features, by Lynn Chua et al.
Training Differentially Private Ad Prediction Models with Semi-Sensitive Features
by Lynn Chua, Qiliang Cui, Badih Ghazi, Charlie Harrison, Pritish Kamath, Walid Krichene, Ravi Kumar, Pasin Manurangsi, Krishna Giri Narra, Amer Sinha, Avinash Varadarajan, Chiyuan Zhang
First submitted to arxiv on: 26 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR); Information Retrieval (cs.IR)
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 In this paper, researchers address the challenge of training machine learning models that balance data privacy with performance. They introduce a new task called “semi-sensitive” feature learning, where some features are publicly available and don’t require protection, while others remain private and must be safeguarded using differential privacy (DP) techniques. The authors propose an algorithm for training DP models in this setting and demonstrate its effectiveness on real-world advertising datasets, outperforming existing methods that either protect all features or only the label. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you’re trying to show ads to people based on their interests. You have some information about each person that’s public, like what they’ve liked online. But you also have sensitive data, like how much money they make, that shouldn’t be shared. The challenge is training a computer program to predict which ads will work best for each person while keeping the sensitive information private. This paper proposes a new way to do this using something called “differential privacy.” It’s like adding noise to the data so it can’t be traced back to any one person. |
Keywords
* Artificial intelligence * Machine learning
