The DRT/LETI/DSYS/SSCE/LSSC laboratory is specialised in signal processing from different sensor types (accelerometer,magnetometer, gyroscope, GPS, Audio, Wi-Fi, Bluetooth, heart rate...) and is concerned with machine learning problems such as automaticrecognition of transportation mode from embedded smartphone sensors, stress assessment from dedicated physiological sensors or sportgesture recognition from a connected watch.
These problems have already been the subject of many research based on traditional classification approaches, i.e. manually constructingthe features used for classification from the available signals.The aim of this thesis is to study if deep learning, a machine learning technique that recently showed outstanding results in imagerecognition, is suited to classification problems with multidimensional time series data.This thesis will benefit from the expertise of LIRIS laboratory in the fields of machine learning and deep learning.
Further details can be found here: http://www-instn.cea.fr/formations/formation-par-la-recherche/doctorat/liste-des-sujets-de-these/deep-learning-applique-a-des-signaux-temporels-multidimensionnels,18-0646.html
17 - rue des Martyrs - 38054 Grenoble Cedex 9
Phone number: +33 4 38 78 51 10
Liris laboratory UMR CNRS 5205
Ecole Centrale de Lyon
36 avenue Guy de Collongue
69134 Ecully Cedex
tel : +33 472 18 6576
fax: +33 472 18 6443
email : firstname.lastname@example.org
(c) GdR 720 ISIS - CNRS - 2011-2018.