Information Theory beyond Communications: Distributed Representations and Deep Learning

Informations projet

STRUDEL

N° de convention de subvention: 797805

DOI

10.3030/797805

Projet clôturé

Date de signature de la CE 5 Avril 2018

Date de début 1 Septembre 2018

Date de fin 31 Août 2020

Financé au titre de

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Coût total

€ 171 349,20

Contribution de l’UE

€ 171 349,20

Coordonné par

CENTRALESUPELEC
France

Objectif

Deep learning is an enormously successful recent paradigm with record-breaking performance in numerous applications. Individual autoencoders (AEs) of a multilayer neural network are trained to convert high-dimensional inputs into low-dimensional codes that allow the reconstruction of the input. Although some explanations appear to be solidly grounded, there is no mathematical understanding of the AE learning process. This project is a collaborative endeavor of researchers with strong complementary backgrounds. Its main innovation is the idea to capitalize on powerful and fertile concepts from information theory (expertise of researcher) in order to advance the state of the art in deep learning (expertise of supervisor at TC). The innovative research work is motivated by our recent insight that there is an intimate relationship between AEs, generative adversarial nets and the information bottleneck method. This method is a model-free approach for extracting information from observed variables that are relevant to hidden representations or labels and will serve as basic building block for an information theory of representation learning. The planned objectives are split into 3 workpackages: 1) information-theoretic criteria and statistical tradeoffs for extracting good representations, 2) structured architectures/algorithms for learning, 3) use of stochastic complexity to assess the descriptive power (model selection) of deep neural networks. Accomplishing the challenging goals of this proposal requires a variety of methodologies with a rich potential for transfer of knowledge between the involved fields of information theory, statistics and machine learning. Our new framework is expected to bridge the gap between theory and practice to facilitate a more thorough understanding and hence improved design of deep learning architectures. The fellow researcher is coordinating the LIA Lab of the CNRS (started in 2017) where he is collaborating with the supervisor at TC

Champ scientifique (EuroSciVoc)

• sciences naturelles › informatique et science de l'information › intelligence artificielle › apprentissage automatique › apprentissage profond

Mots‑clés

• machine learning
• deep learning
• information theory
• autoencoder
• unsupervised learning
• supervised learning
• learning algorithms
• distributed representations
• neural networks

Programme(s)

• H2020-EU.1.3. - EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions PROGRAMME PRINCIPAL
  
• H2020-EU.1.3.2. - Nurturing excellence by means of cross-border and cross-sector mobility
  

Thème(s)

• MSCA-IF-2017 - Individual Fellowships
  

Régime de financement

MSCA-IF - Marie Skłodowska-Curie Individual Fellowships (IF)

Appel à propositions

H2020-MSCA-IF-2017

Coordinateur

CENTRALESUPELEC

• Contribution nette de l'UE: € 171 349,20
• Adresse: PLATEAU DE MOULON 3 RUE JOLIOT CURIE ; 91192 GIF SUR YVETTE ; France
• Type d’activité: Higher or Secondary Education Establishments
• Coût total: € 171 349,20

Partenaires (1)

Partenaire

UNIVERSITE DE MONTREAL

Canada

• Contribution nette de l'UE: € 0,00
• Adresse: CP 6128 Station Centre Ville ; H3C3J7 Montreal
• Type d’activité: Higher or Secondary Education Establishments
• Coût total: € 78 811,20