Abstract Gradient Training: A Unified Certification Framework for Data Poisoning, Unlearning, and Differential Privacy
Preprint
Abstract
The impact of inference-time data perturbation (e.g., adversarial attacks) has been extensively studied in machine learning, leading to well-established certification techniques for adversarial robustness. In contrast, certifying models against training data perturbations remains a relatively under-explored area. These perturbations can arise in three critical contexts: adversarial data poisoning, where an adversary manipulates training samples to corrupt model performance; machine unlearning, which requires certifying model behavior under the removal of specific training data; and differential privacy, where guarantees must be given with respect to substituting individual data points. This work introduces Abstract Gradient Training (AGT), a unified framework for certifying robustness of a given model and training procedure to training data perturbations, including bounded perturbations, the removal of data points, and the addition of new samples. By bounding the reachable set of parameters, i.e., establishing provable parameter-space bounds, AGT provides a formal approach to analyzing the behavior of models trained via first-order optimization methods.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- When Deep Learning Meets Polyhedral Theory: A SurveyINFORMS Journal on Computing, 2026
- Scaling Mixed-Integer Programming for Certification of Neural Network Controllers Using Bounds Tightening2024
- Adversarial Robustness Certification for Bayesian Neural Networks2024
- Poisoning Web-Scale Training Datasets is Practical2024
- Set Propagation Techniques for Reachability AnalysisAnnual Review of Control, Robotics, and Autonomous Systems, 2021
- Strong Mixed-Integer Programming Formulations for Trained Neural Networks2019
- An abstract domain for certifying neural networksProceedings of the ACM on Programming Languages, 2019
- Deep neural networks and mixed integer linear optimizationConstraints, 2018
- Output Reachable Set Estimation and Verification for Multilayer Neural NetworksIEEE Transactions on Neural Networks and Learning Systems, 2018
- Formal Verification of Piece-Wise Linear Feed-Forward Neural Networks2017