Unveiling the Role of Randomization in Multiclass Adversarial Classification: Insights from Graph Theory
AISTATS 2025 multiclass graph theory
262-question exercise bank — revise with hints, or contribute a question.
Paris, FR — teaching maths @ PSL · IBSAI
I am a
Maths teacher at IBSAI · PhD in Computer Science (Dauphine–PSL)
A short intro.
I teach mathematics in the International Bachelor of Science in AI at PSL University.
I did my PhD in Computer Science at Université Paris Dauphine – PSL, supervised by Yann Chevaleyre and Benjamin Negrevergne (LAMSADE).
My dissertation, Understanding the role of randomization in adversarially robust classification, is about using randomness to make ML models a bit more robust.
Courses I've taught:
Papers and preprints, newest first:
AISTATS 2025 multiclass graph theory
ECML 2025 attacks mixtures
NeurIPS 2023 · spotlight spotlight randomization
arXiv preprint 2107.14525 graphs compression
SISAP 2021 · code available graphs
No papers for this year — try another filter.
Small utilities for class:
Tally partial marks live with hotkeys, keep a persistent record, and get mean / median / min / max plus CSV export.
T.02Paste a class list, shuffle it, draw podium winners, split into teams, and track who has been called — with history.
T.03One click, one student. No repeats until everyone has gone. For cold-calling without the guilt.
T.04Balanced groups in one click. Remembers past groups and tries not to repeat pairs.
T.05Full-screen countdown to project. Checkpoints, end bell, overtime count. Survives a refresh.
T.06Shuffle the class into the room plan. Click two seats to swap them.
T.07Plot up to 3 curves live. Drag to pan, scroll to zoom, hover to read values.
T.08Add, multiply, invert. Row-reduction shows each step — handy for live demos.
T.09Paste a column of grades. Histogram, quartiles, pass rate, and a curve hint.
T.10Revision questions drawn from the course TDs. Filter by topic, reveal hints first, check the full solution after.
Email works best.