Tel: +41 (0)27 721 77 41
Fax: +41 (0)27 721 77 12
Idiap Research Institute
Centre du Parc
Rue Marconi 19
PO Box 592
CH – 1920 Martigny
I am a fourth-year Ph.D candidate in electrical engineering at the École Polytechnique Fédérale de Lausanne (EPFL) in Switzerland, under the supervision of Ronan Collobert. Before my Ph.D, I worked as an engineer where I developed different data analysis tools for large-scale applications, such as in genetics.
My main research interests are in natural language processing, AI, and machine learning. I’m particularly focused on building models for large-scale datasets, which led me into deep learning models. As a former engineer I always look for fast systems that work well in practice.
hpca, this toolkit provides an efficient implementation of the Hellinger PCA for computing word embeddings.
Get the source code from GitHub.
Rmixmod, an interface for MIXMOD which provides a tool for supervised and unsupervised classification on qualitative and quantitative data.
This package is available on CRAN.
R. Lebret, D. Grangier, M. Auli. Generating Text from Structured Data with Application to the Biography Domain. Under review. [PDF]
R. Lebret, R. Collobert. “The Sum of Its Parts”: Joint Learning of Word and Phrase Representations with Autoencoders. In ICML Deep Learning Workshop, 2015. [PDF]
R. Lebret, P. O. Pinheiro, R. Collobert. Simple Image Description Generator via a Linear Phrase-Based Model. In ICLR Workshop, 2015. [PDF]
R. Lebret, R. Collobert. N-gram-Based Low-Dimensional Representation for Document Classification. In ICLR Workshop, 2015. [PDF]
R. Lebret, J. Legrand, R. Collobert. Is Deep Learning Really Necessary for Word Embeddings? In NIPS Workshop on Deep Learning, 2013. [PDF]
R. Lebret, S. Iovleff, F. Langrognet, C. Biernacki, G. Celeux, G. Govaert. Rmixmod: The R Package of the Model-Based Unsupervised, Supervised and Semi-Supervised Classification Mixmod Library. In Journal of Statistical Software, 67(1):1–29, 2015. [PDF]