Compromis entre information et calcul

par Laure Aptel

Projet de thèse en Information, communications, électronique

Sous la direction de Aslan Tchamkerten.

Thèses en préparation à Paris Saclay , dans le cadre de Ecole Doctorale de l'Institut Polytechnique de Paris , en partenariat avec LTCI - Laboratoire de Traitement et Communication de l'Information (laboratoire) , ComNum : Communications Numériques (equipe de recherche) et de Télécom Paris (établissement de préparation de la thèse) depuis le 01-10-2017 .

  • Résumé

    Recently, a number of advances in the theory of computation have been made by using information-theoretic arguments ({it{e.g.}}, in communication complexity, streaming, and data structures lower bounds). Conversely, some of the most exciting ongoing work in information theory has focused on problems with a computational component ({it{e.g.}}, distributed function computation, relaying over a network, message-passing algorithms). The goal of this project is to explore the rich interplay between information theory and the theory of computation, and create new connections between both areas. In the first part of the project we will address a fundamental question, motivated by bioinformatics, pertaining to computation over coded data, thereby exploring the link between computation and the information theory of data compression. How to efficiently compress data so that it can easily be queried? In the second part of the project we will address a question motivated by computer networks, thereby exploring the link between computation and the information theory of data transmission. Consider a network of interconnected processors. Processors can perform local computations and can also exchange information with each other. Given a function to be computed by the network, what are the fundamental tradeoffs between the amount of local computations and the amount of exchanged information between processors? This project is expected to shed light on new compression algorithms that allow efficient computation over data and high compression rate. It is also expected to shed light on ways to achieve efficient multi-parti computation.

  • Titre traduit

    Tradeoffs between Information and Computation

  • Résumé

    Voir doc. ci-joint.