Site de Vincent Gripon

Livres

2020

2019

2012

DecisiveNets: Training Deep Associative Memories to Solve Complex Machine Learning Problems

Learning deep representations to solve complex machine learning tasks has become the prominent trend in the past few years. Indeed, Deep Neural Networks are now the golden standard in domains as various as computer vision, natural language processing or even playing combinatorial games. However, problematic limitations are hidden behind this surprising universal capability. Among other things, explainability of the decisions is a major concern, especially since deep neural networks are made up of a very large number of trainable parameters. Moreover, computational complexity can quickly become a problem, especially in contexts constrained by real time or limited resources. Therefore, understanding how information is stored and the impact this storage can have on the system remains a major and open issue. In this chapter, we introduce a method to transform deep neural network models into deep associative memories, with simpler, more explicable and less expensive operations. We show through experiments that these transformations can be done without penalty on predictive performance. The resulting deep associative memories are excellent candidates for artificial intelligence that is easier to theorize and manipulate.

@book{GriLasHac2020,
  author = {Vincent Gripon and Carlos Lassance and
Ghouthi Boukli Hacene},
  editor = {ArXiv Preprint},
  title = {DecisiveNets: Training Deep Associative
Memories to Solve Complex Machine Learning Problems},
  year = {2020},
}

Efficient Representations for Graph and Neural Network Signals

@phdthesis{Gri202012,
  author = {Vincent Gripon},
  title = {Efficient Representations for Graph and
Neural Network Signals},
  school = {ENS Lyon},
  year = {2020},
  month = {December},
}

Uncertainty Principle on Graphs

@inbook{PasGriAlaRab20194,
  author = {Bastien Pasdeloup and Vincent Gripon and
Réda Alami and Michael Rabbat},
  editor = {L. Stankovic and E. Sejdic},
  title = {Uncertainty Principle on Graphs},
  pages = {317--340},
  publisher = {Springer Nature},
  year = {2019},
  series = {Vertex-Frequency Analysis of Graph
Signals},
  month = {April},
}

Petite mathématique du cerveau

Du neurone, composant cérébral fondamental, on sait à peu près tout. De l’information mentale, on ne sait presque rien. Sous quelle forme matérielle notre cerveau range-t-il ses visages connus, ses poèmes et ses numéros de téléphone ? De quelle manière les restitue-t-il ? Ceux qui connaissent bien la biologie du neurone et la neuroanatomie ne sont pas encore en mesure de nous éclairer sur cette problématique de nature purement informationnelle. S’il est indispensable de bien connaître le neurone, cela n’est manifestement pas suffisant pour traiter la question plus spéculative de l’information mentale. D’autres concepts, issus de disciplines scientifiques de prime abord étrangères à la biologie telles que la théorie de l’information et le codage redondant, peuvent aider à trouver des réponses satisfaisantes. Cet ouvrage apporte une première réponse concrète, mathématiquement cohérente et biologiquement plausible, sur la manière dont le réseau neural fixe et remémore ses éléments de connaissance. S’y mêlent, en une théorie originale, neurones et graphes, codes correcteurs d’erreurs et colonnes corticales, messages intemporels et séquences, finalement cliques neurales et autres tournois. Les perspectives de développement offertes par cette théorie et par le modèle de mémoire cérébrale entièrement numérique auquel elle conduit, sont nombreuses et prometteuses, en neurosciences comme dans le champ de l’intelligence artificielle.

This book is currently only available in french.

@book{BerGri201209,
  author = {Claude Berrou and Vincent Gripon},
  editor = {Odile Jacob},
  title = {Petite mathématique du cerveau},
  year = {2012},
  month = {September},
}