Programme
Monday, September 11
13.45 - 14.00: Opening of the workshop
14.00 - 15.30: Set 1
15.30 - 16.00: Coffee break
- Joachim Buhmann: Resilient algorithms by information theory and applications in medicine
- David Bickel: The entropy-maximizing Bayesian model of sufficiently small codelength
- Naftali Tishby: The Information Theory of Deep Learning
16.00 - 16.30: Set 216.30 - 17.30: Plenary Talk: François Baccelli
- Olgica Milenkovic: New Approaches to Hypergraph Clustering and Community Detection
18.00: Welcoming Reception
Tuesday, September 12
09.00 - 09.30: Coffee
09.30 - 10.30: Plenary Talk: Peter Shor
10.30 - 11.00: Coffee Break
11.00 - 12.30: Set 3:
12.30 - 14.00: Lunch
- Alex Jung: Information-Theoretic Limits of Learning Networks for Big Data
- Joe Suzuki: Conditional Mutual Information Estimation and its application to Conditional Independence Detection
- Mérouane Debbah: Random Matrix Methods for Wireless Communications
14.00 - 15.30: Set 4:
15.30 - 16.00: Coffee break
- Peter Harremoës: Improved Tail Bounds related to Information based Testing
- Ugo Vaccaro: Minimum Entropy Joint Distributions with Fixed Marginals: Algorithms and Applications
- Boris Ryabko: Data compressors beyond data compression
16.00 - 17.30: Set 5:
19.30: Banquet Dinner at Alcazar Restaurant (62, rue Mazarine, 75006 Paris)
- Sumio Watanabe: Difference between Bayes Cross Validation and WAIC for Conditional Independent Samples
- Tim van Erven: MetaGrad: Multiple Learning Rates in Online Learning
- Kazuho Watanabe: Rate-distortion dimension and Bayesian learning coefficient
Wednesday, September 13
09.00 - 09.30: Coffee
09.30 - 10.00: Set 5½:
10.00 - 10.40: NEW: Short/Impromptu Presentations Session
- Marek Zaionc: Asymptotic density in logic and computability
10.40 - 11.00: Coffee Break
- Please contact the organizers to check availability
11.00 - 12.30: Set 6:
12.30 - 14.00: Lunch
- Antonio Galves: Context tree models in the brain
- Jean-Marie Gorce: Fundamental limits of bursty multi-user communications
- Prasad Santhanam: Resampling procedures for Markov processes in the slow mixing regime
14.00 - 15.30: Set 7:
15.30 - 16.00: Coffee break
- Kenji Yamanishi: MDL-based gradual model change detection
- Nidhi Hegde: Learning to match in expert systems: experiments on stackexchange data
- Fabien Mathieu: Shannon, Turing and Hats: Information Theory Incompleteness SLIDES
16.00 - 17.00: Set 8:
- Joseph O'Sullivan: Image Reconstruction Applications from an Information Geometry Viewpoint
- Mokshay Madiman: Physical implications of the concentration of information
Plenary speakers
- Information Theory, Quantum Mechanics, and Black Holes
Abstract: The black hole information paradox is the question of whether information which enters a black hole can ever escape. Quantum mechanics says that it must escape when a black hole evaporates. General relativity says that it cannot. Recently, physicists have been looking to information theory to try to resolve this paradox.
Several fundamental modifications have to be made to classical information theory in order to apply it to quantum mechanical systems. Researchers have been studying how to modify it for some time, and we now have a robust theory of quantum information. After introducing the fundamentals of quantum information theory, I will discuss some proposals for how quantum information theory might help resolve the black hole information paradox.
About the speaker: Peter Shor is Morss Professor of Applied Mathematics since 2003. He received the B.A. in mathematics from Caltech in 1981, and the Ph.D. in applied mathematics from MIT in 1985, under the direction of Tom Leighton. Following a postdoctoral fellowship at MSRI, he joined AT&T. He was a member of its Research staff, 1986-2003. He joined the MIT faculty in applied mathematics as full professor in 2003. Professor Shor's research interests are in theoretical computer science: currently on algorithms, quantum computing, computational geometry and combinatorics. In 1998, Peter Shor received the Nevanlinna Prize and the International Quantum Communication Award. He also received the Dickson Prize in Science from Carnegie-Mellon in 1998. He was awarded the Gödel Prize of the ACM and a MacArthur Foundation Fellowship in 1999. He received the King Faisal International Prize in Science in 2002, and was named one of Caltech's Distinguished Alumni in 2007. He is a member of the National Academy of Science (2002), and fellow of the American Academy of Arts and Sciences (2011).
François Baccelli, University of Texas at Austin
- Stochastic Geometry in High Dimension
Abstract: This talk will survey recent results on Boolean stochastic geometry in high dimensional Euclidean spaces.
A Boolean model in $\mathbb{R}^n$ consists of a homogeneous Poisson point process in $\mathbb{R}^n$, and independent and identically distributed random closed sets of $\mathbb{R}^n$ centered on each atom of this point process.
The Shannon regime features a family of Boolean models indexed by $n \ge 1$, where the $n$-th model has a Poisson point process of intensity $e^{n \rho}$ and i.i.d. random compact sets with diameter of order $\sqrt{n}$, with $n$ tending to $\infty$. A typical example is that where each random compact set is an $n$ ball of radius distributed like $\bar X_n \sqrt{n}$, with $\bar X_n$ satisfying a large deviations principle.
The main focus of the talk will be on the asymptotic behavior of classical Boolean stochastic geometry quantities, like volume fraction, percolation threshold or mean cluster size, in this Shannon regime.
More general high dimensional Particle Processes will also be discussed. For instance the case where the point process is hard core or determinantal rather than Poisson, and compact sets are still i.i.d.
This work is motivated by problems in information theory in the Poltyrev regime. The Boolean case corresponds to random coding, and the Particle Process case to more general coding assumptions. It leads to new results on error exponents.
Joint work with V. Anantharam (UC Berkeley), and E. O'Reilly (UT Austin).
About the speaker: François Baccelli is Simons Math+X Chair in Mathematics and ECE at UT Austin. His research directions are at the interface between Applied Mathematics (probability theory, stochastic geometry, dynamical systems) and Communications (network science, information theory, wireless networks). He is co-author of research monographs on point processes and queues (with P. Brémaud); max plus algebras and network dynamics (with G. Cohen, G. Olsder and J.P. Quadrat); stationary queuing networks (with P. Brémaud); stochastic geometry and wireless networks (with B. Blaszczyszyn). Before joining UT Austin, he held positions at INRIA, Ecole Normale Supérieure and Ecole Polytechnique. He received the France Télécom Prize of the French Academy of Sciences in 2002 and the ACM Sigmetrics Achievement Award in 2014. He is a co-recipient of the 2014 Stephen O. Rice Prize and of the Leonard G. Abraham Prize Awards of the IEEE Communications Theory Society. He is a member of the French Academy of Sciences and part time researcher at INRIA.