Information, Privacy and Stability in Adaptive Data Analysis

2017-06-02 · Adam Smith

Traditional statistical theory assumes that the analysis to be performed on a given data set is selected independently of the data themselves. This assumption breaks downs when data are re-used across analyses and the analysis to be performed at a given stage depends on the results of earlier stages. Such dependency can arise when the same data are used by several scientific studies, or when a single analysis consists of multiple stages. How can we draw statistically valid conclusions when data are re-used? This is the focus of a recent and active line of work. At a high level, these results show that limiting the information revealed by earlier stages of analysis controls the bias introduced in later stages by adaptivity. Here we review some known results in this area and highlight the role of information-theoretic concepts, notably several one-shot notions of mutual information.

📄 PDF Abstract BibTeX arXiv:1706.00820

Code (0)

등록된 구현이 없습니다.

Tasks

valid

Optimizer Dynamics at the Edge of Stability with Differential Privacy

2025-12-22 · Ayana Hussain, Ricky Fang arxiv

Deep learning models can reveal sensitive information about individual training examples, and while differential privacy (DP) provides guarantees restricting such leakage, it also alters optimization dynamics in poorly u…

Calibrating Noise to Variance in Adaptive Data Analysis

2017-12-19 · Vitaly Feldman, Thomas Steinke

Datasets are often used multiple times and each successive analysis may depend on the outcome of previous analyses. Standard techniques for ensuring generalization and statistical validity do not account for this adaptiv…

Online Learning via the Differential Privacy Lens

2017-11-27 · NeurIPS 2019 12 · Jacob Abernethy, Young Hun Jung, Chansoo Lee, Audra McMillan 외

In this paper, we use differential privacy as a lens to examine online learning in both full and partial information settings. The differential privacy framework is, at heart, less about privacy and more about algorithmi…

Multi-Armed Bandits

FedCAda: Adaptive Client-Side Optimization for Accelerated and Stable Federated Learning

2024-05-20 · Liuzhi Zhou, Yu He, Kun Zhai, Xiang Liu 외

Federated learning (FL) has emerged as a prominent approach for collaborative training of machine learning models across distributed clients while preserving data privacy. However, the quest to balance acceleration and s…

Federated Learning