A scientific day in honour of Nicolas Sourlas

Location

The conference will be held in the "Conferences 4" hall of the Physics Department of the Ecole Normale Superieure, on friday 6th of june.

To get to the Physics Department, 24 rue Lhomond, see the indications and the maps on the laboratory webpage.

To get to the conference hall : enter the 24 rue Lhomond. Enter the first building on your right, climb up the stairs, then take the elevator to the second floor, take the corridor on your left, then follow the corridor on your right until you reach the "Conferences 4".

You can find a list of hotels nearby the ENS on the webpage http://www.lpt.ens.fr/hotels_list.html.

Program

Registration

There is no registration fee. We however kindly ask you to register if you plan to attend the conference, in order to ease its organization. In particular, we plan to organize a dinner in a restaurant on the friday evening, we hence need to know whether you will join us.

Contact

For further information please contact Jean Iliopoulos, ilio@lpt.ens.fr.

Abstracts of the talks

Giorgio Parisi, An empirical study of large starling flocks

Bird flocking is a striking example of collective animal behaviour. A vivid illustration of this phenomenon is provided by the aerial display of vast flocks of starlings gathering at dusk over the roost and swirling with extraordinary spatial coherence. Both the evolutionary justification and the mechanistic laws of flocking are poorly understood, arguably because of a lack of data on large flocks. Here, I report a quantitative study of aerial display: the individual three-dimensional positions in compact flocks of up to 2700 birds have been measured.

I will discuss some of the main features of the flock as a whole - shape, movement, density and structure. Flocks are relatively thin, with variable sizes, but constant proportions. They tend to slide parallel to the ground and, during turns, their orientation changes with respect to the direction of motion. Individual birds keep a minimum distance from each other that is comparable to their wingspan. The density within the aggregations is non-homogeneous, as birds are packed more tightly at the border compared to the centre of the flock.

These results constitute the first set of large-scale data on three-dimensional animal aggregations. Current models and theories of collective animal behaviour can now be tested against these results.

David Sherrington, A simple spin glass perspective on martensitic shape-memory alloys

Concepts and methodologies developed in the study of spin glasses have led to novel, and often useful, methods to describe, understand and apply complex many body systems in many physically different scenarios. In this talk I shall employ such a perspective to introduce and consider a qualitative mapping for martensitic shape-memory alloys; with the austenite, tweed and twinned phases of the martensitic materials corresponding to paramagnetic, spin glass and periodic phases in pseudo-spin glass alloys.

Hidetoshi Nishimori, Lee-Yang zeros and Griffiths singularities in 2d and 3d spin glasses

The distribution of zeros of the partition function is studied for the two- and three-dimensional Ising spin glasses on the complex field plane, i.e. the Lee-Yang zeros. We estimate the density of zeros on the imaginary-field axis by an importance-sampling Monte Carlo algorithm, which enables us to sample extremely rare events. Our results in 2d clearly suggest that the density has an essential singularity at the origin, giving the first evidence for Griffiths singularities in spin glasses in equilibrium.

Jean-Christian Angles d'Auriac, Random Bond Potts Model and sub-modular functions

An algorithm to compute the free energy of the Random Bond Potts Model at any temperature in the limit of an infinite number of states will be presented. This algorithm works in a time polynomial with the volume of the lattice, for any lattice. On another hand, the ground state of the Random Field Ising Model can also be found in a polynomial time. These two algorithms will be shown to be both inherited from a property of sub-modularity of some function to minimize. To illustrate the method, various results concerning the Random Bond Potts Model will be presented.

Enzo Marinari, Ranking by Loops

I study cycles in sparse random graphs by treating them as a constraint satisfaction problem. I discuss how the model is defined, and the results that can be obtained in this way. In particular, I introduce and discuss an approximation to the self-avoiding walk-sum which emerges naturally from a slightly engineered version of the original problem.

Heinz Horner, Dynamics of of an Ising spin-glass on the Bethe lattice

An Ising-spin-glass model on a Bethe lattice combines aspects of spin glasses in finite dimensions (finite coordination number), as well as in infinite dimensions (absence of loops). This model has also ben investigated in the context of satisfiability problems.

At present Glauber dynamics for Ising spins with binary random couplings $J_{ij}=\pm J_o$ is investigated. A continuous transition to a non ergodic low temperature state is found. An expansion around the critical temperature shows that this system is in the universality class of the SK-model with mutual interactions of all spins. The nature of the low temperature phase for general distributions $P(J_{ij})$ is not of this type, but no other satisfactory solution could be found so far. Investigations within replica theory have been restricted to binary couplings only, leaving the case of more general distributions open as well.

Jean-Pierre Nadal, Neural coding: an information theoretic approach

In computational neuroscience, information theory is widely used for quantifying the efficiency of neural processing. I will illustratethis on the particular case of "population coding", a kind of neural representation which is observed in both sensory and motor areas of the cortex: a large population of neurons is recruted for representing (coding), e.g., a direction ('head direction cells' in the rat), a location ('place cells'), etc. I will present some results obtained on the optimal coding efficiency of such neural codes.

David Forney, From Spin Glass Models to Codes on Graphs

Ruediger Urbanke, Nicalos Sourlas

In 1989 N. Sourlas described and analyzed low-density generator (LDGM) codes as spin systems. Since LDGM codes necessarily exhibit high error floors, further studies of this class of codes were initially limited.

The picture has changed significantly in recent years, LDGM codes can be seen as predecessors of rate-less codes and they form the basis for the source coding problem. After a quick review I will look at the simplest lossy source coding problem, namely the compression of a binary symmetric source using LDGM codes. I will discuss the biggest remaining challenge -- a rigorous analysis of the iterative decoding algorithm, which is based on a sequence of message-passing and decimation steps. [Based on joint work with S. Korada and S. Kudekar.]

Sergio Verdu, Divergence

Jorge Kurchan, Mapping non-equilibrium into equilibrium

The fluctuations of out of equilibrium systems are much richer than those of equilibrium ones because time-reversibility is lost, and the Gibbs-Boltzmann distribution does not hold.

A general theory for large, smooth (`hydrodynamic') deviations has been developed in the last few years: it offers us a glimpse of what an out of equilibrium thermodynamics might one day look like.

When applied to several simple yet non-trivial one-dimensional transport models, it was found that a complete analytic solution can be obtained. It turns that at the bottom of this miracle is the fact that these systems have the unexpected feature that a non-local tranformation maps the driven version back into an equilibrium one.