Hi, I'm Mohammed-Yassine
PhD student in interpretable machine learning for ecology — I study how deep representations of biodiversity data can be aligned with ecological concepts.
- Interpretability
- Concept learning
- Ecology & biodiversity
- Self-supervised learning
- Adversarial robustness
I am a PhD student at Inria Grenoble in the Thoth team and at the Laboratoire d’Écologie Alpine (LECA), Université Grenoble Alpes, supervised by Karteek Alahari, Hadrien Hendrikx and Sara Si-Moussi. My thesis, Aligning deep learning with ecological concepts: explainable methods for biodiversity modelling, asks how ecological interactions can be inferred from the latent representations of deep models trained on multimodal biodiversity data (images, sound, species observations).
I also work on robust and self-supervised representation learning inspired by neuroscience — oscillatory neurons and predictive learning — a line of research started at OIST (Japan) with Prof. Makoto Yamada. Before that, I was a research assistant at Columbia University (USA).
I am a civil servant engineer of the French Corps des ingénieurs des ponts, des eaux et des forêts (IPEF) and trained in applied mathematics and machine learning at École polytechnique (X2021) and in the MVA master at ENS Paris-Saclay.
News
- Sep 2026Started my PhD at Inria Grenoble (Thoth team) and LECA on explainable deep learning for biodiversity modelling.
- Sep 2026New preprint on arXiv: Neither Adversarial Training Nor Purification: Emergent Adversarial Robustness from Oscillatory Predictive Learning.
- Feb 2026Research internship at I4CE (Institute for Climate Economics) on international transition finance.
- Sep 2025Joined the Corps des IPEF as an Ingénieur-élève des ponts, des eaux et des forêts and started the PAPDD specialised master at École des Ponts & AgroParisTech.
- Sep 2025Our paper Neuroscience-Inspired Encoding and Learning: A Path to Robust Representation Learning was accepted at the WiML workshop @ NeurIPS 2025.
- Aug 2025Completed the MVA master (Mathematics, Vision, Learning) at ENS Paris-Saclay.
Research interests
Publications
Neither Adversarial Training Nor Purification: Emergent Adversarial Robustness from Oscillatory Predictive Learning
arXiv preprint arXiv:2609.08683, 2026
- Adversarial robustness
- Self-supervised learning
- Oscillatory neurons
Neuroscience-Inspired Encoding and Learning: A Path to Robust Representation Learning
WiML Workshop @ NeurIPS 2025, 2025
- Adversarial robustness
- Self-supervised learning
- Neuroscience-inspired
All publications → · All projects →
Experience
- Sep 2026 – presentPhD Student Inria Grenoble, Thoth team & LECA, Université Grenoble Alpes Explainable deep learning for biodiversity modelling, supervised by Karteek Alahari, Hadrien Hendrikx and Sara Si-Moussi. Project page.
- Feb – Jul 2026Research Intern I4CE — Institute for Climate Economics, Paris, France Comparative analysis of financing strategies and structural mechanisms to catalyse investment in the real economy of developing countries.
- Sep 2025 – presentCivil servant engineer, Corps des IPEF Ingénieurs des ponts, des eaux et des forêts — French Ministries of Agriculture and Ecological Transition My PhD is part of the doctoral training track of the corps.
- Apr – Aug 2025Research Assistant OIST, Machine Learning and Data Science Unit — Okinawa, Japan Neuroscience-inspired encoding and learning for robust representation learning, supervised by Prof. Makoto Yamada. Work accepted at the WiML workshop @ NeurIPS 2025.
- Apr – Jul 2024Research Assistant Columbia University, Irving Institute for Cancer Dynamics — New York, USA Deep learning and random-forest Approximate Bayesian Computation to infer the parameters of finite-sites evolutionary models from cancer genomic data, supervised by Dr. Simon Tavaré. Project page.
- Jun – Aug 2023Application Support Intern Eurofins — Hamburg, Germany Analysed error logs, maintained the backlog and built KPI dashboards for the Eurofins network of analysis laboratories.
- 2022 – 2023Teaching Assistant in Mathematics Lycée Sainte-Geneviève — Versailles, France Advised and evaluated students preparing for the entrance exams of French engineering schools.
Education
- 2026 – 2029PhD in interpretable machine learning for ecology Université Grenoble Alpes — doctoral school MSTII Aligning deep learning with ecological concepts: explainable methods for biodiversity modelling.
- 2025 – 2026Specialised Master PAPDD — Public Policies and Actions for Sustainable Development École nationale des ponts et chaussées & AgroParisTech (Université Paris-Saclay) Law, markets and regulation, public management, environmental assessment, ecological transition, agriculture and food, energy, forests.
- 2024 – 2025Master MVA — Mathematics, Vision, Learning ENS Paris-Saclay (Université Paris-Saclay) Representation learning, geometric deep learning, LLMs, optimal transport, kernel methods, reinforcement learning, inverse problems, generative models.
- 2021 – 2024Ingénieur polytechnicien program (X2021) — Master's degree in applied mathematics École polytechnique (Institut Polytechnique de Paris) Stochastic calculus, Monte Carlo methods, optimisation and control, statistical learning theory, random models for biology and ecology.
- 2019 – 2021Classes préparatoires (mathematics, physics, computer science) Lycée Sainte-Geneviève — Versailles, France
Beyond research
🎻 I have played the violin since childhood (8 years of conservatory, 10 years of orchestra across four different orchestras) and I also sing in a choir — I am always happy to discover new musical repertoires. In my free time I enjoy running and cycling, or improving my skills at rhythm games such as DanceDanceRevolution.
🌍 Languages: French (native), English (proficient), Chinese and Arabic (intermediate).