Many
studies have examined the risk factors for relapse in alcohol-dependent
patients within the first year of treatment, and have generally focused on two
personality dimensions: emotional instability and difficulty in establishing
relationships. In this study, we examine if these weaknesses remain in
prolonged alcohol abstinence. To do so, we administer the NEO PI-R to two
groups of subjects. Group 1, Inactive Drinkers (ID), consists of 51 patients
with at least two years of abstinence (average length of abstinence for this
group is 6.2 years); Group 2, Recently Detoxified Drinkers (RDD), comprises 93
patients who have recently ceased consuming alcohol. Based on NEO PI-R scores,
our results evidence that inactive drinkers experience significant reduction in
emotional instability and improvement in relationships to others. We further
observe that, with long-term abstinence, these personality dimensions
normalize, ceasing to be risk factors for relapse. Additionally, we find that
this long-term amelioration of traits altered by alcohol amounts to an improved
behavioral adaptation to life events rather than an actual change in personality.

Abstract:
We performed a community research program in order to analyze the evolution of interpersonal values following alcohol withdrawal in alcoholics attending to a self-help group. The Gordon questionnaire on interpersonal values was administered every 3 months during one year to 145 individuals having recently stopped drinking. At baseline, scores of 5 interpersonal categories (dependence, conformism, consideration, independence, kindness) were in the medium interval of usual values while that of command category was low, specifically in men. Values of conformism and independence increased according to time while those of dependence and consideration decreased but the differences were significant only in those who remained abstinent all along the observation period. In those who relapsed, there were no modifications; however, sharing these values might have allowed these subjects to be less isolated. We will describe in details the nature of these evolutions.

Abstract:
We show that near a point where the equilibrium density of eigenvalues of a matrix model behaves like y ~ x^{p/q}, the correlation functions of a random matrix, are, to leading order in the appropriate scaling, given by determinants of the universal (p,q)-minimal models kernels. Those (p,q) kernels are written in terms of functions solutions of a linear equation of order q, with polynomial coefficients of degree at most p. For example, near a regular edge y ~ x^{1/2}, the (1,2) kernel is the Airy kernel and we recover the Airy law. Those kernels are associated to the (p,q) minimal model, i.e. the (p,q) reduction of the KP hierarchy solution of the string equation. Here we consider only the 1-matrix model, for which q=2.

Abstract:
We compute the mixed correlation function in a way which involves only the orthogonal polynomials with degrees close to $n$, (in some sense like the Christoffel Darboux theorem for non-mixed correlation functions). We also derive new representations for the differential systems satisfied by the biorthogonal polynomials, and we find new formulae for the spectral curve. In particular we prove the conjecture of M. Bertola, claiming that the spectral curve is the same curve which appears in the loop equations.

Abstract:
We use techniques from relative algebraic geometry and homotopical algebraic geometry in order to construct several categories of schemes defined "under Spec Z". We define this way the categories of N-schemes, F_1-schemes, S-schemes, S_+-schemes, and S_1-schemes, where from a very intuitive point of view N is the semi-ring of natural numbers, F_1 is the field with one element, S is the sphere ring spectrum, S_+ is the semi-ring spectrum of natural numbers and S_1 is the ring spectrum with one element. These categories of schemes are related by several base change functors, and they all possess a base change functor to Z-schemes (in the usual sense). Finally, we show how the linear group Gl_n and toric varieties can be defined as objects in certain of these categories.

Abstract:
Gaussian mixture models are widely used to study clustering problems. These model-based clustering methods require an accurate estimation of the unknown data density by Gaussian mixtures. In Maugis and Michel (2009), a penalized maximum likelihood estimator is proposed for automatically selecting the number of mixture components. In the present paper, a collection of univariate densities whose logarithm is locally {\beta}-H\"older with moment and tail conditions are considered. We show that this penalized estimator is minimax adaptive to the {\beta} regularity of such densities in the Hellinger sense.

Abstract:
This paper is about variable selection, clustering and estimation in an unsupervised high-dimensional setting. Our approach is based on fitting constrained Gaussian mixture models, where we learn the number of clusters $K$ and the set of relevant variables $S$ using a generalized Bayesian posterior with a sparsity inducing prior. We prove a sparsity oracle inequality which shows that this procedure selects the optimal parameters $K$ and $S$. This procedure is implemented using a Metropolis-Hastings algorithm, based on a clustering-oriented greedy proposal, which makes the convergence to the posterior very fast.

Abstract:
We find new representations for Itzykson-Zuber like angular integrals for arbitrary beta, in particular for the orthogonal group O(n), the unitary group U(n) and the symplectic group Sp(2n). We rewrite the Haar measure integral, as a flat Lebesge measure integral, and we deduce some recursion formula on n. The same methods gives also the Shatashvili's type moments. Finally we prove that, in agreement with Brezin and Hikami's observation, the angular integrals are linear combinations of exponentials whose coefficients are polynomials in the reduced variables (x_i-x_j)(y_i-y_j).

Abstract:
We prove that the correlations functions, generated by the determinantal process of the Christoffel-Darboux kernel of an arbitrary order 2 ODE, do satisfy loop equations.

Abstract:
This work is concerned with the analysis of a stochastic approximation algorithm for the simulation of quasi-stationary distributions on finite state spaces. This is a generalization of a method introduced by Aldous, Flannery and Palacios. It is shown that the asymptotic behavior of the empirical occupation measure of this process is precisely related to the asymptotic behavior of some deterministic dynamical system induced by a vector field on the unit simplex. This approach provides new proof of convergence as well as precise rates for this type of algorithm. We then compare this algorithm with particle system algorithms.