Research topics
My research lies at the crossroads of machine learning, signal processing, applied mathematics and data mining. This theme covers a wide range of work on the analysis and processing of signals and time series, with applications in medicine, biology and industry.
Event detection in time series
|
★ Change-point detection: parametric and nonparametric techniques, supervised approaches, applications...
★ Anomaly detection: unsupervised approaches, topological data analysis...
★ Pattern and motif discovery: similarity search, elastic distances, shape analysis...
|
|
Representation learning for time series
|
★ Dictionary learning and sparse coding: convolutional sparse coding, tensor approaches, graph dictionary learning...
★ Symbolization for time series: adaptive representations, distances between multivariate time series...
★ Unsupervised representation learning with deep learning
|
Graph signal processing for multivariate time series
|
★ Graph signal processing: graph learning, dictionary learning, filtering, sampling, denoising, interpolation, sensor networks...
★ Network Granger causality estimation
|
|
Interdisciplinarity: applications to medicine, biology and industry
|
★ Biomedical signal processing: accelerometry data, 2D and 3D trajectories (eye tracking, posture and motion capture), ECG, EEG, respiratory signals, video, sound...
★ Application to biomedical research and neurosciences: gait analysis, posture analysis, motion analysis, oculomotricity, plethysmography, general anesthesia, cardiology, mental workload, behavior analysis...
★ Industrial applications: predictive maintenance
|
Reproducibility and open science
|
★ Open-source software and librairies
★ Online web applications and reproducible publications
★ Publication of medical databases
More info on the Software page
|
|
