The first part of this thesis introduces new algorithms for the sparse encoding of signals. Based on Matching Pursuit (MP) they focus on the following problem : how to reduce the computation time of the selection step of MP. As an answer, we sub-sample the dictionary in line and column at each iteration. We show that this theoretically grounded approach has good empirical performances. We then propose a bloc coordinate gradient descent algorithm for feature selection problems in the multiclass classification setting. Thanks to the use of error-correcting output codes, this task can be seen as a simultaneous sparse encoding of signals problem. The second part exposes new empirical Bernstein inequalities. Firstly, they concern the theory of t...
This thesis presents various problems of adaptive functional estimation, using projection and kernel...
In the parametric estimation context, estimators performances can be characterized, inter alia, by t...
Many algorithmic problems are « hard », in the sense of we do not know how to solve them in polynomi...
The first part of this thesis introduces new algorithms for the sparse encoding of signals. Based on...
Classical SAR Imagery algorithms or SAR processors are all based on the isotropic point model. When ...
Learning stochastic models generating sequences has many applications in natural language processing...
This thesis is concerned with the estimation of the dynamical parameters of one or multiple targets ...
In computer science, a lot of applications use distances. In the context of structured data, strings...
In computer science, a lot of applications use distances. In the context of structured data, strings...
The thesis objective is to develop a procedure for the calibration of LHCb's flavour tagging algorit...
The thesis objective is to develop a procedure for the calibration of LHCb's flavour tagging algorit...
In multiobjective optimization, the knowledge of the Pareto set provides valuable information on the...
The thesis objective is to develop a procedure for the calibration of LHCb's flavour tagging algorit...
This thesis deals with the problem of global optimization of expensive-to-evaluate functions in a Ba...
This thesis studies the following topics in Machine Learning: Bandit theory, Statistical learning an...
This thesis presents various problems of adaptive functional estimation, using projection and kernel...
In the parametric estimation context, estimators performances can be characterized, inter alia, by t...
Many algorithmic problems are « hard », in the sense of we do not know how to solve them in polynomi...
The first part of this thesis introduces new algorithms for the sparse encoding of signals. Based on...
Classical SAR Imagery algorithms or SAR processors are all based on the isotropic point model. When ...
Learning stochastic models generating sequences has many applications in natural language processing...
This thesis is concerned with the estimation of the dynamical parameters of one or multiple targets ...
In computer science, a lot of applications use distances. In the context of structured data, strings...
In computer science, a lot of applications use distances. In the context of structured data, strings...
The thesis objective is to develop a procedure for the calibration of LHCb's flavour tagging algorit...
The thesis objective is to develop a procedure for the calibration of LHCb's flavour tagging algorit...
In multiobjective optimization, the knowledge of the Pareto set provides valuable information on the...
The thesis objective is to develop a procedure for the calibration of LHCb's flavour tagging algorit...
This thesis deals with the problem of global optimization of expensive-to-evaluate functions in a Ba...
This thesis studies the following topics in Machine Learning: Bandit theory, Statistical learning an...
This thesis presents various problems of adaptive functional estimation, using projection and kernel...
In the parametric estimation context, estimators performances can be characterized, inter alia, by t...
Many algorithmic problems are « hard », in the sense of we do not know how to solve them in polynomi...