Consider Glauber dynamics for the Ising model on the hypercubic lattice with a positive magnetic field. Starting from the minus configuration, the system initially settles into a metastable state with negative magnetization. Slowly the system relaxes to a stable state with positive magnetization. Schonmann and Shlosman showed that in the two dimensional case the relaxation time is a simple function of the energy required to create a critical Wulff droplet. The dilute Ising model is obtained from the regular Ising model by deleting a fraction of the edges of the underlying graph. In this paper we show that even an arbitrarily small dilution can dramatically reduce the relaxation time. This is because of a catalytic effect—rare regions of high dilution speed up the transition from minus phase to plus phase.

In this paper we consider the Glauber dynamics for a disordered ferromagnetic Ising model, in the region of phase coexistence. It was conjectured several decades ago that the spin autocorrelation decays as a negative power of time [Huse and Fisher, in Phys Rev B 35(13):6841–6846, 1987 ]. We confirm this behavior by establishing a corresponding lower bound in any dimensions d ≥ 2, together with an upper bound when d = 2. Our approach is deeply connected to the Wulff construction for the dilute Ising model. We consider initial phase profiles with a reduced surface tension on their boundary and prove that, under mild conditions, those profiles are separated from the (equilibrium) pure plus phase by an energy barrier.

Motivated by a recent use of Glauber dynamics for Monte Carlo simulations of path integral representation of quantum spin models (Krzakala et al. in Phys. Rev. B 78(13):134428, 2008 ), we analyse a natural Glauber dynamics for the quantum Ising model with a transverse field on a finite graph G . We establish strict monotonicity properties of the equilibrium distribution and we extend (and improve) the censoring inequality of Peres and Winkler to the quantum setting. Then we consider the case when G is a regular b -ary tree and prove the same fast mixing results established in Martinelli et al. (Commun. Math. Phys. 250(2):301–334, 2004 ) for the classical Ising model. Our main tool is an inductive relation between conditional marginals (known as the “cavity equation”) together with sharp bounds on the operator norm of the derivative at the stable fixed point. It is here that the main difference between the quantum and the classical case appear, as the cavity equation is formulated here in an infinite dimensional vector space, whereas in the classical case marginals belong to a one-dimensional space.

We study a quenched charged-polymer model, introduced by Garel and Orland in 1988 (Europhys. Lett. 6(7):597–601, 1988 ; Europhys. Lett. 6(4):307–310, 1988 ), that reproduces the folding/unfolding transition of biopolymers. We prove that, below the critical inverse temperature, the polymer is delocalized in the sense that: (1) the rescaled trajectory of the polymer converges to the Brownian path; and (2) the partition function remains bounded. At the critical inverse temperature, we show that the maximum time spent at points jumps discontinuously from 0 to a positive fraction of the number of monomers, in the limit as the number of monomers tends to infinity. Finally, when the inverse temperature is large, we prove that the polymer collapses in the sense that a large fraction of its monomers live on four adjacent positions, and its diameter grows only logarithmically with the number of the monomers. Our methods also provide some insight into the annealed phase transition and at the transition due to a pulling force; both phase transitions are shown to be discontinuous.

We study a quenched charged-polymer model, introduced by Garel and Orland in 1988, that reproduces the folding/unfolding transition of biopolymers. We prove that, below the critical inverse temperature, the polymer is delocalized in the sense that: (1) The rescaled trajectory of the polymer converges to the Brownian path; and (2) The partition function remains bounded.

The Curie-Weiss Potts model is a mean field version of the well-known Potts model. In this model, the critical line beta = beta(c)(h) is explicitly known and corresponds to a first-order transition when q > 2. In the present paper we describe the fluctuations of the density vector in the whole domain beta >= 0 and h >= 0, including the conditional fluctuations on the critical line and the non-Gaussian fluctuations at the extremity of the critical line. The probabilities of each of the two thermodynamically stable states on the critical line are also computed. Similar results are inferred for the random-cluster model on the complete graph. (C) 2009 Published by Elsevier B.V.

We study the surface tension and the phenomenon of phase coexistence for the Ising model on \({\mathbb{Z}^d\,(d\, \geqslant\, 2)}\) with ferromagnetic but random couplings. We prove the convergence in probability (with respect to random couplings) of surface tension and analyze its large deviations: upper deviations occur at volume order while lower deviations occur at surface order. We study the asymptotics of surface tension at low temperatures and relate the quenched value τ q of surface tension to maximal flows (first passage times if d =  2). For a broad class of distributions of the couplings we show that the inequality \({\tau^a\, \leqslant\, \tau^q}\) –where τ a is the surface tension under the averaged Gibbs measure – is strict at low temperatures. We also describe the phenomenon of phase coexistence in the dilute Ising model and discuss some of the consequences of the media randomness. All of our results hold as well for the dilute Potts and random cluster models.

We study the Kertész line of the q-state Potts model at (inverse) temperature β in the presence of an external magnetic field h. This line separates the two regions of the phase diagram according to the existence or not of an infinite cluster in the Fortuin–Kasteleyn representation of the model. It is known that the Kertész line hK(β) coincides with the line of first order phase transition for small fields when q is large enough. Here, we prove that the first order phase transition implies a jump in the density of the infinite cluster; hence, the Kertész line remains below the line of first order phase transition. We also analyze the region of large fields and prove, using techniques of stochastic comparisons, that hK(β) equals log(q−1)−log(β−βp) to the leading order, as β goes to βp=−log(1−pc), where pc is the threshold for bond percolation.

We present a study of phase transitions of the mean-field Potts model at (inverse) temperature β, in the presence of an external field h. Both thermodynamic and topological aspects of these transitions are considered. For the first aspect we complement previous results and give an explicit equation of the thermodynamic transition line in the β-h plane as well as the magnitude of the jump of the magnetization (for ). The signature of the latter aspect is characterized here by the presence or not of a giant component in the clusters of a Fortuin-Kasteleyn type representation of the model. We give the equation of the Kertész line separating (in the β-h plane) the two behaviours. As a result, we get that this line exhibits, as soon as , a very interesting cusp where it separates from the thermodynamic transition line.

By means of a multi-scale analysis we describe the typical geometrical structure of the clusters under the FK measure in random media. Our result holds in any dimension greater or equal to 2 provided that slab percolation occurs under the averaged measure, which should be the case in the whole supercritical phase. This work extends the one of Pisztora and provides an essential tool for the analysis of the supercritical regime in disordered FK models and in the corresponding disordered Ising and Potts models.

We present a study of phase transitions of the Curie--Weiss Potts model at (inverse) temperature $\beta$, in presence of an external field $h$. Both thermodynamic and topological aspects of these transitions are considered. For the first aspect we complement previous results and give an explicit equation of the thermodynamic transition line in the $\beta$--$h$ plane as well as the magnitude of the jump of the magnetization (for $q \geqslant 3)$. The signature of the latter aspect is characterized here by the presence or not of a giant component in the clusters of a Fortuin--Kasteleyn type representation of the model. We give the equation of the Kert\'esz line separating (in the $\beta$--$h$ plane) the two behaviours. As a result, we get that this line exhibits, as soon as $q \geqslant 3$, a very interesting cusp where it separates from the thermodynamic transition line.

Digital Cellular telephony requires voice compression designed to minimize the bandwidth required for the digital cellular channel. The features used in speech recognition have similar components to those used in the vocoding process. The present invention provides a system that bypasses the de-compression or decoding phase of the vocoding and converts the digital cellular parameters directly into features that can be processed by a recognition engine. More specifically, the present invention provides a system and method for mapping a vocoded representation of parameters defining speech components, which in turn define a particular waveform, into a base feature type representation of parameters defining speech components (e.g. LPC parameters), which in turn define the same digital waveform.