Board and administration

Directeur : Jean Ponce

Secrétariat : Isabelle Delais

Tél. : +33 (0)1 44 32 20 45 - Fax : +33 (0)1 44 32 20 75

Adresse : Ecole normale supérieure, 45 rue d’Ulm, 75230 Paris Cedex 05



37 permanent professors and researchers
60 PhD students, Post-Docs, and temporary researchers

  • Antique — Static analysis by abstract interpretation (head: Xavier Rival)
  • Cascade — Cryptography (head: David Pointcheval)
  • Data — Signal Processing and Classification (head: Stéphane Mallat)
  • Dyogene — Dynamics of Geometric Networks (head: Marc Lelarge)
  • Parkas — Parallelism of Synchronous Kahn Networks (head: Marc Pouzet)
  • Sierra — Machine Learning (head: Francis Bach)
  • Talgo — Theory, Algorithms, topoLogy, Graphs, and Optimization (head: Claire Mathieu)
  • Willow — Artificial Vision (head: Jean Ponce)

Research themes

Static analysis by abstract interpretation:

The Antique CNRS/ENS/INRIA Team carries out research in semantics, static analysis, abstract interpretation of programs and biological systems. We search for automatic techniques to compute semantic properties of programs (in particular safety critical embedded programs) and biological systems. Properties of interest include, among others, safety (such as absence of runtime errors, preservation of data invariants), liveness (such as termination), security (absence of information leak). Such properties are usually not computable, or can only be computed at a prohibitive cost in the case of finite systems. Therefore we tackle them with techniques based on conservative abstraction.


The research activity of the project-team CASCADE addresses the following topics, which cover almost all the domains that are currently active in the international cryptographic community, but mainly in the public-key area:
- Design and Provable Security in Public-Key Cryptography
- Randomness in Cryptography
- Lattice Cryptography
- Security amidst Concurrency on the Internet

Signal Processing and Classification:

- Invariant Representations with Scattering
- Sparse Dictionary and Unsupervised Group Learning
- Data Geometry
- Inverse Problems 

Dynamics of Geometric Networks:

- Perfect simulation
- Stochastic geometry and information theory
- The cavity method for network algorithms
- Statistical learning
- Network calculus

Parallelism of Synchronous Kahn Networks: 

Formally defined high-level languages for embedded systems 
- design, semantics and implementation of programming languages
- synchronous data-flow concurrency
- address new applications: computationally intensive, large-scale simulation, mix of continuous/discrete time

Efficient compilation for Modern architectures 
- internal and formally defined representations of optimizing compilers (e.g., polyhedral compilation algorithms)
- generate provably correct and efficient code from synchronous designs
- target modern shared-memory parallel processors (e.g., multi/many cores, tiled processors arrays, GPUs)

Machine Learning:

- Supervised learning
- Unsupervised learning
- Parsimony
- Optimization

Theory, Algorithms, topoLogy, Graphs, and Optimization:

The current research themes of the team include the design of algorithms and study of combinatorial structures for geometrical and topological problems, and for combinatorial optimization. We use tools from algebraic topology, combinatorics, probability theory, and mathematical programming. In terms of communities, the main relevant fields are discrete algorithms (SODA) and computational geometry (SoCG), with links to combinatorics and graph theory. The following four research directions are currently investigated:
- algorithms for embedded graphs: approximation algorithms for network design problems on planar graphs; algorithms for topological problems in graphs on surfaces;
- approximation algorithms and techniques for combinatorial optimization;
- streaming and online algorithms for graph problems;
- combinatorial geometry problems with a topological flavor.

Artificial Vision:

- 3D object and scene modeling, analysis, and retrieval
- Human activity capture and classification
- Category-level object and scene recognition
- Machine learning

Département d’Informatique de l’ENS

UMR 8548 Ecole Normale Supérieure