arXiv subjects

P. L. Garrido

Publications and source records attributed to P. L. Garrido.

28 records · Page 2Linked to original sources

arXiv2005

Effects of fast presynaptic noise in attractor neural networks

We study both analytically and numerically the effect of presynaptic noise on the transmission of information in attractor neural networks. The noise occurs on a very short-time scale compared to that for the neuron dynamics and it produces short-time synaptic depression. This is inspired in recent neurobiological findings that show that synaptic strength may either increase or decrease on a short-time scale depending on presynaptic activity. We thus describe a mechanism by which fast presynaptic noise enhances the neural network sensitivity to an external stimulus. The reason for this is that, in general, the presynaptic noise induces nonequilibrium behavior and, consequently, the space of fixed points is qualitatively modified in such a way that the system can easily scape from the attractor. As a result, the model shows, in addition to pattern recognition, class identification and categorization, which may be relevant to the understanding of some of the brain complex tasks.

q-bio.NC↗

arXiv2004

Theoretical Characterization of the Interface in a Nonequilibrium Lattice System

The influence of nonequilibrium bulk conditions on the properties of the interfaces exhibited by a kinetic Ising--like model system with nonequilibrium steady states is studied. The system is maintained out of equilibrium by perturbing the familiar spin--flip dynamics at temperature T with completely--random flips; one may interpret these as ideally simulating some (dynamic) impurities. We find evidence that, in the present case, the nonequilibrium mechanism adds to the basic thermal one resulting on a renormalization of microscopic parameters such as the probability of interfacial broken bonds. On this assumption, we develop theory for the nonequilibrium "surface tension", which happens to show a non--monotonous behavior with a maximum at some finite T. It ensues, in full agreement with Monte Carlo simulations, that interface fluctuations differ qualitatively from the equilibrium case, e.g., the interface remains rough at zero--T. We discuss on some consequences of these facts for nucleation theory, and make some explicit predictions concerning the nonequilibrium droplet structure.

cond-mat.stat-mech↗

arXiv2004

Reentrant Behavior of the Spinodal Curve in a Nonequilibrium Ferromagnet

The metastable behavior of a kinetic Ising--like ferromagnetic model system in which a generic type of microscopic disorder induces nonequilibrium steady states is studied by computer simulation and a mean--field approach. We pay attention, in particular, to the spinodal curve or intrinsic coercive field that separates the metastable region from the unstable one. We find that, under strong nonequilibrium conditions, this exhibits reentrant behavior as a function of temperature. That is, metastability does not happen in this regime for both low and high temperatures, but instead emerges for intermediate temperature, as a consequence of the non-linear interplay between thermal and nonequilibrium fluctuations. We argue that this behavior, which is in contrast with equilibrium phenomenology and could occur in actual impure specimens, might be related to the presence of an effective multiplicative noise in the system.

cond-mat.stat-mech↗

arXiv2003

Effects of Static and Dynamic Disorder on the Performance of Neural Automata

We report on both analytical and numerical results concerning stochastic Hopfield--like neural automata exhibiting the following (biologically inspired) features: (1) Neurons and synapses evolve in time as in contact with respective baths at different temperatures. (2) The connectivity between neurons may be tuned from full connection to high random dilution or to the case of networks with the small--world property and/or scale-free architecture. (3) There is synaptic kinetics simulating repeated scanning of the stored patterns. Though these features may apparently result in additional disorder, the model exhibits, for a wide range of parameter values, an extraordinary computational performance, and some of the qualitative behaviors observed in natural systems. In particular, we illustrate here very efficient and robust associative memory, and jumping between pattern attractors.

cond-mat.stat-mech↗

arXiv2003

The Boltzmann Entropy for Dense Fluids Not in Local Equilibrium

We investigate, via computer simulations, the time evolution of the (Boltzmann) entropy of a dense fluid not in local equilibrium. The macrovariables $M$ describing the system are the (empirical) particle density $f=\{f(\un{x},\un{v})\}$ and the total energy $E$. We find that $S(f_t,E)$ is monotone increasing in time even when its kinetic part is decreasing. We argue that for isolated Hamiltonian systems monotonicity of $S(M_t) = S(M_{X_t})$ should hold generally for ``typical'' (the overwhelming majority of) initial microstates (phase-points) $X_0$ belonging to the initial macrostate $M_0$, satisfying $M_{X_0} = M_0$. This is a direct consequence of Liouville's theorem when $M_t$ evolves autonomously.

cond-mat.stat-mech↗

arXiv2003

Influence of topology on the performance of a neural network

We studied the computational properties of an attractor neural network (ANN) with different network topologies. Though fully connected neural networks exhibit, in general, a good performance, they are biologically unrealistic, as it is unlikely that natural evolution leads to such a large connectivity. We demonstrate that, at finite temperature, the capacity to store and retrieve binary patterns is higher for ANN with scale--free (SF) topology than for highly random--diluted Hopfield networks with the same number of synapses. We also show that, at zero temperature, the relative performance of the SF network increases with increasing values of the distribution power-law exponent. Some consequences and possible applications of our findings are discussed.

cond-mat.stat-mech↗

arXiv2000

Universality classes of driven lattice gases

Motivated by some recent criticisms to our alternative Langevin equation for driven lattice gases (DLG) under an infinitely large driving field, we revisit the derivation of such an equation, and test its validity. As a result, an additional term, coming from a careful consideration of entropic contributions, is added to the equation. This term heals all the recently reported generic infrared singularities. The emerging equation is then identical to that describing randomly driven diffusive systems. This fact confirms our claim that the infinite driving limit is singular, and that the main relevant ingredient determining the critical behavior of the DLG in this limit is the anisotropy and not the presence of a current. Different aspects of our picture are discussed, and it is concluded that it constitutes a very plausible scenario to rationalize the critical behavior of the DLG and variants of it.

cond-mat↗

arXiv1998

Mesoscopic description of the annealed Ising model and Multiplicative noise

A new type of Langevin equation exhibiting a non trivial phase transition associated with the presence of multiplicative noise is introduced. The equation is derived as a mesoscopic representation of the microscopic annealed Ising model (AIM) proposed by Thorpe and Beeman, and reproduces perfectly its basic phenomenology. The AIM exhibits a non-trivial behavior as the temperature is increased, in particular it presents a disorder-to-order phase transition at low temperatures, and a order-to-disorder transition at higher temperatures. This behavior resembles that of some Langevin equations with multiplicative noise, which exhibit also two analogous phase transitions as the noise-amplitude is increased. By comparing the standard models for noise-induced transitions with our new Langevin equation we elucidate that the mechanisms controlling the disorder-to-order transitions in both of them are essentially different, even though for both of them the presence of multiplicative noise is a key ingredient.

cond-mat↗

arXiv1998

Continuum Field Description of Driven Lattice Gases

We review the critical behaviour of the Driven Lattice Gas (DLG) model. As a result, we obtain a novel Langevin equation for the DLG which depends on the microscopic transition probabilities. We then show how this dependence affects the critical behaviour of the the DLG placing the finite and the infinite driving field cases into different universality classes. Two other well known anisotropic, conservative, non-equilibrium models, the two-temperature model and the randomly driven model (RDLG) are also studied. It is shown that the RDLG with infinite averaged field and the two-temperature model with infinite parallel temperature fall in the same universality class as the infinitely driven DLG. A Langevin equation for the two-layer DLG is also presented.

cond-mat↗

arXiv1996

Chaotic principle: an experimental test

The chaotic hypothesis discussed in [GC1] is tested experimentally in a simple conduction model. Besides a confirmation of the hypothesis predictions the results suggest the validity of the hypothesis in the much wider context in which, as the forcing strength grows, the attractor ceases to be an Anosov system and becomes an Axiom A attractor. A first test of the new predictions is also attempted.

chao-dyn↗