Machine-checked statistical learning theory in Lean 4: empirical-Bernstein and time-uniform PAC-Bayes, Markov risk, Rademacher/VC, and Dudley chaining.
-
Updated
Sep 4, 2026 - Lean
Machine-checked statistical learning theory in Lean 4: empirical-Bernstein and time-uniform PAC-Bayes, Markov risk, Rademacher/VC, and Dudley chaining.
Implementation of PDFAs and PDFA learning algorithm.
Machine Learning for Data 3141 Reichman University Spring 2022 - 6 Homework Projects
🔨 A prototype tool on learning DOTAs by testing.
My solutions to the assignments in Introduction to Machine Learning course at Tel Aviv University (course number: 0368-3235)
Learning Reliable Rules under Class Imbalance (SDM 2021)
Machine Learning and Analysis of Big Data course, Computer Science M.Sc., Ben Gurion University of the Negev, 2020
This project dives into mathematical foundations of PAC learning theory while adding Python examples that make core ideas like generalization, sample complexity, VC dimension, halfspaces, perceptrons, and logical semantics executable and easier to inspect.
Homework Advanced Machine Learning (AML), MSc Artificial Intelligence, Year 1, Semester 2, Faculty of Mathematics and Computer Science, University of Bucharest
Code for Bounded Information: PAC Certification of Multivariate Side-Channel Traces (ASIACRYPT 2026). Computes certified bounds, in bits, on side-channel leakage for a declared set of attackers, with a Docker package that rebuilds every table and figure of the paper.
To associate your repository with the pac-learning topic, visit your repo's landing page and select "manage topics."