FÉLIX CHAVELLI

Hi there! I am Félix Chavelli, a second year Ph.D. student in Computer Science at Inria Paris and École normale supérieure - PSL University, within the VALDA team. My research focuses on representation and interpretability in (multivariate) time series. I am supervised by Paul Boniol and Michael Thomazo.

Prior to this, I completed a double M.Sc. program in Computer Science at the National University of Singapore and in Applied Mathematics at ENSTA Paris - Institut Polytechnique de Paris. I previously earned a B.Sc. in Applied Mathematics at Université de Toulouse.

PROJECTS

• Okinawa Institute of Science and Technology | Development of interpretable methods for time series segmentation, supervised by Prof. Makoto Yamada at OIST - Machine Learning and Data Science unit. Japan, June - August 2025
• European Space Agency | Analysis of ESA Climate Change Initiative contribution to IPCC climate science reports. Harwell, UK, February - May 2024
• CNRS | Physics-informed neural networks for dynamical systems at CNRS@CREATE (Centre National de la Recherche Scientifique) in Singapore, under the supervision of Stéphane Bressan, September 2022 to November 2023
• IPCC | Meta-analysis of the 6th Assessment Report of the IPCC (Intergovernmental Panel on Climate Change), under the supervision of Sarah Connors, September 2021 to February 2022
• Harvard University | Physics-informed neural networks for Navier-Stokes equations in turbulent channel flow, within the StellarDNN team at John A. Paulson School of Engineering and Applied Sciences (SEAS), under the supervision of David Sondak and Pavlos Protopapas, May - August 2021
• IRIT | Machine learning for predictive maintenance of aircraft engines within the SAMoVA team of IRIT (Toulouse Institute for Research in Computer Science) in collaboration with ISAE Supaéro, under the supervision of Thomas Pellegrini, June 2019
• Météo France | Statistical modeling applied to visibility and fog phenomena at the Forecasting Operations Department, under the supervision of Olivier Mestre, August 2018

PUBLICATIONS

• NeurIPS Felix Chavelli, Paul Boniol, Michaël Thomazo. Toward Interpretable Evaluation Measures for Time Series Segmentation. In Proceedings of the Thirty-Ninth Conference on Neural Information Processing Systems (NeurIPS 2025), San Diego, USA, December 2025.
• VLDBBDA EGC Felix Chavelli. Toward Interpretable Methods for Time Series Analytics. 51st International Conference on Very Large Data Bases (VLDB 2025) PhD Workshop, London, United Kingdom | 41èmes journées de la conférence BDA « Gestion de Données – Principes, Technologies et Applications » (PhD Workshop), Toulouse, France | Vers des mesures d’évaluation interprétables pour la segmentation de séries temporelles. 26ème Conférence Francophone sur l'Extraction et la Gestion des Connaissances (EGC 2026), Anglet, France
• NeurIPS Félix Chavelli, Zi-Yu Khoo, Dawen Wu, Jonathan Sze Choong Low, Stéphane Bressan. Physics-informed Discovery of State Variables in Second-Order and Hamiltonian Systems. Machine Learning and the Physical Sciences Workshop, NeurIPS 2024, Vancouver, BC, Canada
• ACIIDS Félix Chavelli, Zi-Yu Khoo, Dawen Wu, Jonathan Sze Choong Low, Stéphane Bressan. Physics-informed Discovery of State Variables in Second-Order and Hamiltonian Systems. Intelligent Information and Database Systems, ACIIDS 2025. In: Nguyen, N.T., et al. Intelligent Information and Database Systems. ACIIDS 2025. Lecture Notes in Computer Science, vol 15683. Springer, Singapore.
• DEXA Félix Chavelli, Zi-Yu Khoo, Dawen Wu, Jonathan Sze Choong Low, Stéphane Bressan. Assessing the Effectiveness of Intrinsic Dimension Estimators for Uncovering the Phase Space Dimensionality of Dynamical Systems from State Observations. Database and Expert Systems Applications. DEXA 2023. Lecture Notes in Computer Science, vol 14146. Springer, Cham.

TEACHING

• Lecturer | PSL University - International Bachelor of Science in AI, The Art of Computer Programming, 2026 (Semester 2)
• Teaching Assistant | PSL University & Lycée Louis-le-Grand - CPES – Sciences des données, arts et cultures, Programming, 2025 (Semester 1)