Réunion

Geometry and time-series signal processing 

Annonce

2 November 2026 – Paris 

In many signal processing and time series applications (navigation, motion analysis, sensor networks, etc.), measured data evolve in structured, non-necessarily Euclidean spaces. Over the past few years, a new research trend has emerged, incorporating these geometric constraints (Lie groups, Riemannian manifolds, etc.) into data processing methods (change-point detection, filtering, etc.). In parallel, tools from algebraic geometry and topology (signature theory, topological data analysis) have shed new light on time series analysis. The purpose of this scientific workshop is to present the latest advances regarding the links between signal processing and geometry (in a broad sense).

The topics include (not limited)

• Signal processing on manifolds
• Change-point detection on non-Euclidean space
• Signature theory and rough path
• Topological data analysis (TDA)
• Statistical learning on geometric spaces
• Analysis of structured time series

Invited speakers (confirmed) 

• Fréderic Chazal (INRIA Saclay)
• Marianne Clausel (CRAN/University of Lorraine)
• Karim Dahia (ONERA)

Call for contributions

Early-career researchers (Ph.D. students, postdocs, recently appointed assistant professors) are encouraged to apply. Please send your title and a short abstract (max. 1/2 page) to samy.labsir@ipsa.fr, laurent.oudre@ens-paris-saclay.fr , hugo.henneuse@univ-evry.fr,  by October 2.
Decisions on oral or poster presentations will be sent by October 9.

Organizers

• Samy Labsir (IPSA/TéSA–ISAE-SUPAERO)
• Laurent Oudre (Centre Borelli, ENS Paris-Saclay)
• Hugo Henneuse (Université Evry Paris-Saclay)