Talks
My conference presentations and poster sessions at academic venues.
Conference Presentations
Michelin DoctoBib'Day 2026
A Unified Online Framework for Adaptive Soft Sensing in High-Dimensional Batch Processes
April 13, 2026 · Clermont-Ferrand, France
Louis Berthier1,2, Ahmed Shokry1, Maxime Moreaud2, Guillaume Ramelet2, Eric Moulines1
1 CMAP, École Polytechnique 2 Michelin, Clermont-Ferrand
Predicting product quality in real time is hard when the manufacturing process itself keeps drifting. This work benchmarks five adaptive soft sensing strategies, from temporal weighting to neighborhood-based retrieval, leveraging Self-Organizing Maps (SOMs) with online SHAP-driven feature selection (SHAP scores how much each variable drives a prediction). Evaluated on 35,125 production batches and 167 process variables at Michelin, it gives practical guidance for choosing the right online adaptation strategy in industrial settings.
ESCAPE 35
Knowledge Discovery in Large-Scale Batch Processes through Explainable Boosted Models and Uncertainty Quantification: Application to Rubber Mixing
July 9, 2025 · Ghent, Belgium
Louis Berthier1,2, Ahmed Shokry1,*, Eric Moulines1, Sylvain Desroziers1, Guillaume Ramelet2
1 CMAP, CNRS, École Polytechnique, IP Paris 2 Michelin
Rubber compounding involves hundreds of interacting process variables, and only a few actually drive product quality. This work takes an explainability-first approach, combining gradient-boosted trees (a strong tabular model built from many small decision trees) with SHAP attribution and conformal prediction, giving engineers both interpretable insights and statistically rigorous uncertainty estimates. The result: process experts can pinpoint critical quality drivers with quantified confidence. Presented at ESCAPE 35, one of Europe's premier conferences in computer-aided chemical engineering.
Posters
Local Dynamic Calibration via Self-Organized CP Presented at 2 venues
SIAM UQ26 · Minneapolis, USA · March 22-25, 2026
Michelin DoctoBib'Day 2026 · Clermont-Ferrand, France · April 13, 2026
Louis Berthier, Ahmed Shokry, Maxime Moreaud, Guillaume Ramelet, Aymeric Dieuleveut
Standard conformal prediction gives prediction intervals that are correct *on average*, which can hide undercoverage in specific, safety-critical process regimes. Self-Organized CP closes that gap by discovering groups directly from the input geometry with a Self-Organizing Map (SOM), then calibrating locally inside each one, with no supervised labels and no model retraining. I presented this poster at two venues: SIAM UQ26, a leading conference on uncertainty quantification for industrial applications, and Michelin's annual doctoral day.
Knowledge Discovery in Large-Scale Batch Processes
December 2024 · Paris, France · Welcome Day IP Paris
Louis Berthier, Ahmed Shokry, Sylvain Desroziers, Guillaume Ramelet, Eric Moulines
A black-box quality prediction is only useful once a process engineer can act on it. This poster presents a unified framework pairing gradient boosted models with SHAP attribution and conformal coverage guarantees, translating raw predictions into interpretable, uncertainty-aware process insights at production scale.
Awarded Best Poster in the Mathematics category.
A Framework for Knowledge Discovery in Rubber Mixing Processes
November 2024 · Clermont-Ferrand, France · Michelin Doctoral Day
Louis Berthier, Ahmed Shokry, Sylvain Desroziers, Guillaume Ramelet, Eric Moulines
The first iteration of my knowledge discovery framework, focused on offline analysis. Gradient boosted regression combined with SHAP attribution and conformal prediction intervals surfaces the most influential process variables and their interactions, giving rubber compounding engineers a clear, quantified view of what drives product quality.
2DSBG: A 2D Semi Bi-Gaussian Filter for Line Feature Detection
June 2023 · Rhodes, Greece · ICASSP 2023
Louis Berthier, Adrien Ruggiero, Marcel Pie, Ghulam Sakhi Shokouh, Baptiste Magnier
Thin line features in noisy images are notoriously hard to detect with sub-pixel accuracy. The 2DSBG filter uses an asymmetric Gaussian kernel that selectively enhances elongated structures while suppressing background noise, outperforming classical symmetric approaches like the Laplacian-of-Gaussian on both synthetic and real-world benchmarks. Presented at ICASSP, the flagship IEEE conference on signal processing.