From 2012 to 2015, I was a Ph.D. student in the SIERRA team, part of both Inria and École Normale Supérieure computer science department. I was advised by Sylvain Arlot and Francis Bach. My thesis work focused on machine learning for structured signals (images, audio signals and others kind of sequential signals). More precisely I worked about how to learn similarity measures for algorithms with complex outputs. I defended on September, 18th, 2015.

This webpage is not updated anymore. I have a new webpage as well as a backup of this one.

Publications

A Weakly-Supervised Discriminative Model for Audio-to-score Alignment
Rémi Lajugie, Piotr Bojanowski, Philippe Cuvillier, Sylvain Arlot, Francis Bach.
In Proc. International Conference on Accoustics, Speech and Signal Processing, 2016.
Instance-level video segmentation from object tracks
Proc. CVPR, 2016.
Prédiction structurée pour l'analyse de données séquentielles. Structured prediction for sequential data
Tel Archives ouvertes.
Weakly-Supervised Alignment of Video With Text
Proc. International Conference on Computer Vision, 2015.
Semidefinite and Spectral Relaxations for Multi-Label Classification
Technical report (hal-01159321), 2015.
Learning the Metric for Aligning Temporal Sequences
In Proc. Neural Information Processing Systems 2014.
* = authors contributed equally

Weakly Supervised Action Labeling in Videos Under Ordering Constraints
In Proc. European Conference on Computer Vision 2014.
Large Margin Metric Learning for Constrained Partitioning Problems
In Proc. International Conference on Machine Learning 2014.

Teaching

Cours d'apprentissage statistique, L3 course, filière maths-informatique, École Normale Supérieure, Teaching assistant, 2013.
Cours d'apprentissage statistique, M1 course, filière maths-informatique, École Normale Supérieure, Teaching assistant, 2014.