Tom Pelletreau-Duris

PhD candidate in graph machine learning at the Vrije Universiteit Amsterdam under the supervision of Dr. Jieying Cheng and Pr. Michael Cochez, with Pr. Stefan Scholbach as advisor. Part of the KAI research group. Guest in the L&R group. Half funded by the Zorro project. A consortium with ASML, Canon Production Printers, ITEC, Philips, and ThermoFisher Scientific.

Officially advancing explainability of neuro-symbolic AI. Tactically working on mechanistic interpretability for graph ml. Personally trying to bridge disciplinaries I love, networking philosophy of mind, social sciences, AI and neurosciences through the lens of graph theory and XAI. Systematizing rhizomes.

Tom Pelletreau-Duris

Interactive CV Network

Latest posts

The Docile Learner

2026-04-22 · foucault, deleuze, canguilhem, education, machine learning, knowledge tracing, AI, neurodivergence, philosophy

There is a scene that repeats itself, at this hour, in several million browser tabs. A learner sits before a screen. The screen asks a question. The learner answers. Somewhere — in a server that nobody has ever visited, in a Dutch or Irish data centre cooled by the North Sea — a probability is updated. A number that did not exist a moment ago now exists, and it stands for the learner. The number is small: a scalar between zero and one, or a vector of a few dozen dimensions. But it is, for the duration of the session, who the learner is to the machine. The next question will be selected on its basis. The next hint, withheld or granted, on its basis. The next suggested lesson, the next notification, the congratulatory animation of a green owl or a red flame — all of it flows from this number.

Latest publication

Do Graph Neural Network States Contain Graph Properties?

Tom Pelletreau-Duris, Ruud van Bakel, Michael Cochez — Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning ·

Presents a model-agnostic explainability pipeline that uses diagnostic classifiers to probe graph neural network embeddings for structural graph properties, clarifying what information their hidden states retain across architectures and datasets.

© Pelletreau-Duris Tom — 2025
Built with Nuxt.js