Skip to content

Certification-Based Differentially Private Learning

Sep 2026 · arXiv.org · Vol abs/2609.39629 · 1 citation · 56 references
Computer Science

TL;DR

This work generalizes the abstract gradient training (AGT) framework and introduces Abstract Gradient Sampling (AGS), an algorithm that enables reachability-based analysis to provide guarantees for private learning and finds that under matched conditions, the private learning algorithm can outperform standard private learners.

Abstract

Differential privacy (DP) in machine learning is typically achieved by adding noise to model parameters (private learning) or to model outputs (private prediction). Recent work uses formal methods, namely abstract interpretation, to provide tighter privacy guarantees, but only for private prediction in classification settings. In this work, we investigate the use of formal methods as a general tool for tighter privacy analysis. First, we generalize the abstract gradient training (AGT) framework to private prediction in continuous, unbounded regression. Second, by reducing learning in parameterized models to a regression problem over the parameter space, we introduce Abstract Gradient Sampling (AGS), an algorithm that enables reachability-based analysis to provide guarantees for private learning. In both private prediction and private learning, we provide tightened privacy accounting for the AGT framework and a theoretical analysis demonstrating when our smooth sensitivity upper-bounds yield favourable privacy-utility trade-off. In practice, we validate that our regression bounds are tighter than global-sensitivity baselines on regression benchmarks, and, notably, yield the first finite privacy guarantees in settings where global prediction sensitivity is a priori unbounded. We also find that under matched conditions, our private learning algorithm can outperform standard private learners.

View source

Similar papers

#machine learning Preprint Oct 2026

Certification-Enhanced Generalization Bounds

We investigate the use of formal methods to provide tight and sound generalization bounds for learning algorithms. By casting the traditional notion of algorithmic stability as a specification to be verified, we demonstrate that recent advances in reachability analysis can yield provable bounds on the generalization of...

Leo Elmecker-Plakolm, Matthew Wicker · 0 citations
Preprint Oct 2026

Complexity of Differentially Private Selection via Federated APIs

Differential privacy (DP) has become the standard for privacy-preserving data analysis, yet its practical implementation in federated or distributed settings often faces a significant utility gap compared to the central model. While existing frameworks, including local, shuffle, and secure aggregation models, offer vet...

Hilal Asi, V. Feldman, Jelani Nelson et al. · 0 citations
#large language models Open access Oct 2026

Differential Privacy Tradeoffs in Large-Scale ML Model Training

This paper introduces DPOptimizer, a privacy budget allocation framework that maximizes model utility for a given privacy budget through adaptive noise calibration, layer-wise privacy amplification, and optimal composition strategies, and provides the first systematic analysis of how the privacy-utility-compute trilemm...

S. Soni, Rinku Vedu Patili, Pritam Samanta et al. · 0 citations

OPFL: Optimistic Verification of Federated Learning via Empirical Boundary

OPFL calibrates an empirical boundary offline and uses it to distinguish benign numerical deviations from malicious manipulation, and adopts optimistic verification by post auditing only sampled training steps to reduce the cost of expensive MPC replay.

Hong-Xu Su, Jian-Zhu Yao, Xue-Chao Wang et al. · 0 citations

Related blog posts

MIT News · Artificial Intelligence Nov 17, 2026

Learning the language of machine learning

By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.

Microsoft Research Blog Nov 16, 2026

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.