-
Expansion of net correlations in terms of partial correlations
Authors:
Bautista Arenaza,
Sebastián Risau-Gusman,
Inés Samengo
Abstract:
Graphical models are usually employed to represent statistical relationships between pairs of variables when all the remaining variables are fixed. In this picture, conditionally independent pairs are disconnected. In the real world, however, strict conditional independence is almost impossible to prove. Here we use a weaker version of the concept of graphical models, in which only the linear comp…
▽ More
Graphical models are usually employed to represent statistical relationships between pairs of variables when all the remaining variables are fixed. In this picture, conditionally independent pairs are disconnected. In the real world, however, strict conditional independence is almost impossible to prove. Here we use a weaker version of the concept of graphical models, in which only the linear component of the conditional dependencies is represented. This notion enables us to relate the marginal Pearson correlation coefficient (a measure of linear marginal dependence) with the partial correlations (a measure of linear conditional dependence). Specifically, we use the graphical model to express the marginal Pearson correlation $ρ_{ij}$ between variables $X_i$ and $X_j$ as a sum of the efficacies with which messages propagate along all the paths connecting the variables in the graph. The expansion is convergent, and provides a mechanistic interpretation of how global correlations arise from local interactions. Moreover, by weighing the relevance of each path and of each intermediate node, an intuitive way to imagine interventions is enabled, revealing for example what happens when a given edge is pruned, or the weight of an edge is modified. The expansion is also useful to construct minimal equivalent models, in which latent variables are introduced to replace a larger number of marginalised variables. In addition, the expansion yields an alternative algorithm to calculate marginal Pearson correlations, particularly beneficial when partial correlation matrix inversion is difficult. Finally, for Gaussian variables, the mutual information is also related to message-passing efficacies along paths in the graph.
△ Less
Submitted 2 April, 2024;
originally announced April 2024.
-
Sleep quality, chronotype and social jet lag of adolescents from a population with a very late chronotype
Authors:
Andrés H. Calderón,
Romina A. Capellino,
Damián Dellavale,
D. Lorena Franco,
Pablo M. Gleiser,
Sergio Lindenbaum,
Mara López-Wortzman,
Sabrina C. Riva,
Fernanda R. Román,
Sebastián Risau-Gusman
Abstract:
Sleep disorders can be a negative factor both for learning as for the mental and physical development of adolescents. It has been shown that, in many populations, adolescents tend to have a poor sleep quality, and a very late chronotype. Furthermore, these features peak at adolescence, in the sense that adults tend to sleep better and have an earlier chronotype. But what happens when we consider a…
▽ More
Sleep disorders can be a negative factor both for learning as for the mental and physical development of adolescents. It has been shown that, in many populations, adolescents tend to have a poor sleep quality, and a very late chronotype. Furthermore, these features peak at adolescence, in the sense that adults tend to sleep better and have an earlier chronotype. But what happens when we consider adolescents in a population where already adults have poor sleep quality and a very late chronotype? We have conducted two non-clinical studies in the city of Bariloche, Argentina aimed at measuring sleep quality, chronotype, and social jet lag, using the Pittsburgh and Munich questionnaires. These were administered individually to groups of high school students, as well as to smaller samples of adults and preadolescents, in order to study differences between adolescents and these groups. The results show that in this population sleep quality is much poorer than in most other healthy populations recorded elsewhere. Furthermore, sleep quality is consistently worse for adolescents than for the other groups. The difference with adults seems to be due mainly to increased daytime sleepiness and sleep latency, whereas the difference with preadolescents seems to be due mainly to shorter sleep duration. We also found that the chronotypes of all the groups are very late, with a peak at an age between 18 and 24 ys. Social jet lag and sleep onset latency are also large, and they peak at adolescence, which suggests that they might be closely related to the large prevalence of poor sleep quality that we find in adolescents.
△ Less
Submitted 19 March, 2021;
originally announced March 2021.
-
Backbone structure of the Edwards-Anderson spin-glass model
Authors:
F. Roma,
S. Risau-Gusman
Abstract:
We study the ground-state spatial heterogeneities of the Edwards-Anderson spin-glass model with both bimodal and Gaussian bond distributions. We characterize these heterogeneities by using a general definition of bond rigidity, which allows us to classify the bonds of the system into two sets, the backbone and its complement, with very different properties. This generalizes to continuous distribut…
▽ More
We study the ground-state spatial heterogeneities of the Edwards-Anderson spin-glass model with both bimodal and Gaussian bond distributions. We characterize these heterogeneities by using a general definition of bond rigidity, which allows us to classify the bonds of the system into two sets, the backbone and its complement, with very different properties. This generalizes to continuous distributions of bonds the well known definition of a backbone for discrete bond distributions. By extensive numerical simulations we find that the topological structure of the backbone for a given lattice dimensionality is very similar for both discrete and continuous bond distributions. We then analyze how these heterogeneities influence the equilibrium properties at finite temperature and we discuss the possibility that a suitable backbone picture can be relevant to describe spin-glass phenomena.
△ Less
Submitted 4 October, 2013; v1 submitted 5 April, 2013;
originally announced April 2013.
-
Intransitivity and coexistence in four species cyclic games
Authors:
Alessandra F. Lütz,
Sebastián Risau-Gusman,
Jeferson J. Arenzon
Abstract:
Intransitivity is a property of connected, oriented graphs representing species interactions that may drive their coexistence even in the presence of competition, the standard example being the three species Rock-Paper-Scissors game. We consider here a generalization with four species, the minimum number of species allowing other interactions beyond the single loop (one predator, one prey). We sho…
▽ More
Intransitivity is a property of connected, oriented graphs representing species interactions that may drive their coexistence even in the presence of competition, the standard example being the three species Rock-Paper-Scissors game. We consider here a generalization with four species, the minimum number of species allowing other interactions beyond the single loop (one predator, one prey). We show that, contrary to the mean field prediction, on a square lattice the model presents a transition, as the parameter setting the rate at which one species invades another changes, from a coexistence to a state in which one species gets extinct. Such a dependence on the invasion rates shows that the interaction graph structure alone is not enough to predict the outcome of such models. In addition, different invasion rates permit to tune the level of transitiveness, indicating that for the coexistence of all species to persist, there must be a minimum amount of intransitivity.
△ Less
Submitted 18 February, 2013; v1 submitted 29 May, 2012;
originally announced May 2012.
-
Levy ratchets with dichotomic random flashing
Authors:
S. A. Ibáñez,
A. B. Kolton,
S. Risau-Gusman,
S. Bouzat
Abstract:
Additive symmetric Lévy noise can induce directed transport of overdamped particles in a static asymmetric potential. We study, numerically and analytically, the effect of an additional dichotomous random flashing in such Lévy ratchet system. For this purpose we analyze and solve the corresponding fractional Fokker-Planck equations and we check the results with Langevin simulations. We study the b…
▽ More
Additive symmetric Lévy noise can induce directed transport of overdamped particles in a static asymmetric potential. We study, numerically and analytically, the effect of an additional dichotomous random flashing in such Lévy ratchet system. For this purpose we analyze and solve the corresponding fractional Fokker-Planck equations and we check the results with Langevin simulations. We study the behavior of the current as function of the stability index of the Lévy noise, the noise intensity and the flashing parameters. We find that flashing allows both to enhance and diminish in a broad range the static Lévy ratchet current, depending on the frequencies and asymmetry of the multiplicative dichotomous noise, and on the additive Lévy noise parameters. Our results thus extend those for dichotomous flashing ratchets with Gaussian noise to the case of broadly distributed noises.
△ Less
Submitted 2 September, 2011;
originally announced September 2011.
-
Of lice and math: using models to understand and control populations of head lice
Authors:
María Fabiana Laguna,
Sebastián Risau-Gusman
Abstract:
In this paper we use detailed data about the biology of the head louse (pediculus humanus capitis) to build a model of the evolution of head lice colonies. Using theory and computer simulations, we show that the model can be used to assess the impact of the various strategies usually applied to eradicate head lice, both conscious (treatments) and unconscious (grooming). In the case of treatments,…
▽ More
In this paper we use detailed data about the biology of the head louse (pediculus humanus capitis) to build a model of the evolution of head lice colonies. Using theory and computer simulations, we show that the model can be used to assess the impact of the various strategies usually applied to eradicate head lice, both conscious (treatments) and unconscious (grooming). In the case of treatments, we study the difference in performance that arises when they are applied in systematic and non-systematic ways. Using some reasonable simplifying assumptions (as random mixing of human groups and the same mobility for all life stages of head lice other than eggs) we model the contagion of pediculosis using only one additional parameter. It is shown that this parameter can be tuned to obtain collective infestations whose variables are compatible with what is given in the literature on real infestations. We analyze two scenarios: one where group members begin treatment when a similar number of lice are present in each head, and another where there is one individual who starts treatment with a much larger threshold ('superspreader'). For both cases we assess the impact of several collective strategies of treatment.
△ Less
Submitted 28 April, 2011; v1 submitted 25 April, 2011;
originally announced April 2011.
-
The topological issues of cooperation
Authors:
M. N. Kuperman,
S. Risau-Gusman
Abstract:
In the last years the Prisoner Dilemma (PD) has become a paradigm for the study of the emergence of cooperation in spatially structured populations. Such structure is usually assumed to be given by a graph. In general, the success of cooperative strategies is associated with the possibility of forming globular clusters, which in turn depends on a feature of the network that is measured by its clus…
▽ More
In the last years the Prisoner Dilemma (PD) has become a paradigm for the study of the emergence of cooperation in spatially structured populations. Such structure is usually assumed to be given by a graph. In general, the success of cooperative strategies is associated with the possibility of forming globular clusters, which in turn depends on a feature of the network that is measured by its clustering coefficient. In this work we test the dependence of the success of cooperation with the clustering coefficient of the network, for several different families of networks. We have found that this dependence is far from trivial. Additionally, for both stochastic and deterministic dynamics we have also found that there is a strong dependence on the initial composition of the population. This hints at the existence of several different mechanisms that could promote or hinder cluster expansion. We have studied in detail some of these mechanisms by concentrating on completely ordered networks (large clustering coefficient) or completely random networks (vanishing clustering coefficient).
△ Less
Submitted 25 March, 2011;
originally announced March 2011.
-
Ground-state topology of the Edwards-Anderson +/-J spin glass model
Authors:
F. Roma,
S. Risau-Gusman,
A. J. Ramirez-Pastor,
F. Nieto,
E. E. Vogel
Abstract:
In the Edwards-Anderson model of spin glasses with a bimodal distribution of bonds, the degeneracy of the ground state allows one to define a structure called backbone, which can be characterized by the rigid lattice (RL), consisting of the bonds that retain their frustration (or lack of it) in all ground states. In this work we have performed a detailed numerical study of the properties of the RL…
▽ More
In the Edwards-Anderson model of spin glasses with a bimodal distribution of bonds, the degeneracy of the ground state allows one to define a structure called backbone, which can be characterized by the rigid lattice (RL), consisting of the bonds that retain their frustration (or lack of it) in all ground states. In this work we have performed a detailed numerical study of the properties of the RL, both in two-dimensional (2D) and three-dimensional (3D) lattices. Whereas in 3D we find strong evidence for percolation in the thermodynamic limit, in 2D our results indicate that the most probable scenario is that the RL does not percolate. On the other hand, both in 2D and 3D we find that frustration is very unevenly distributed. Frustration is much lower in the RL than in its complement. Using equilibrium simulations we observe that this property can be found even above the critical temperature. This leads us to propose that the RL should share many properties of ferromagnetic models, an idea that recently has also been proposed in other contexts. We also suggest a preliminary generalization of the definition of backbone for systems with continuous distributions of bonds, and we argue that the study of this structure could be useful for a better understanding of the low temperature phase of those frustrated models.
△ Less
Submitted 9 December, 2010; v1 submitted 6 August, 2010;
originally announced August 2010.
-
Influence of network dynamics on the spread of sexually transmitted diseases
Authors:
Sebastian Risau-Gusman
Abstract:
Network epidemiology often assumes that the relationships defining the social network of a population are static. The dynamics of relationships is only taken indirectly into account, by assuming that the relevant information to study epidemic spread is encoded in the network obtained by considering numbers of partners accumulated over periods of time roughly proportional to the infectious period o…
▽ More
Network epidemiology often assumes that the relationships defining the social network of a population are static. The dynamics of relationships is only taken indirectly into account, by assuming that the relevant information to study epidemic spread is encoded in the network obtained by considering numbers of partners accumulated over periods of time roughly proportional to the infectious period of the disease at hand. On the other hand, models explicitly including social dynamics are often too schematic to provide a reasonable representation of a real population, or so detailed that no general conclusions can be drawn from them. Here we present a model of social dynamics that is general enough that its parameters can be obtained by fitting data from surveys about sexual behaviour, but that can still be studied analytically, using mean field techniques. This allows us to obtain some general results about epidemic spreading. We show that using accumulated network data to estimate the static epidemic threshold leads to a significant underestimation of it. We also show that, for a dynamic network, the relative epidemic threshold is an increasing function of the infectious period of the disease, implying that the static value is a lower bound to the real threshold.
△ Less
Submitted 8 April, 2010;
originally announced April 2010.
-
Do the right thing
Authors:
M. F. Laguna,
G. Abramson,
S. Risau-Gusman,
J. R. Iglesias
Abstract:
We study a model of opinion formation where the opinions in conflict are not equivalent. This is the case when the subject of the decision is to respect a norm or a law. In such scenarios, one of the possible behaviors is to abide by the norm and the other to ignore it. The evolution of the dynamics is implemented through an imitation mechanism, in which agents can change their opinions based on t…
▽ More
We study a model of opinion formation where the opinions in conflict are not equivalent. This is the case when the subject of the decision is to respect a norm or a law. In such scenarios, one of the possible behaviors is to abide by the norm and the other to ignore it. The evolution of the dynamics is implemented through an imitation mechanism, in which agents can change their opinions based on the opinions of a set of partners and their own state. We determine, for different social situations, the minimum percentage of supporters of the law necessary to arrive at a state of consensus of law abiders.
△ Less
Submitted 26 March, 2010;
originally announced March 2010.
-
Contact switching as a control strategy for epidemic outbreaks
Authors:
Sebastian Risau-Gusman,
Damian H. Zanette
Abstract:
We study the effects of switching social contacts as a strategy to control epidemic outbreaks. Connections between susceptible and infective individuals can be broken by either individual, and then reconnected to a randomly chosen member of the population. It is assumed that the reconnecting individual has no previous information on the epidemiological condition of the new contact. We show that…
▽ More
We study the effects of switching social contacts as a strategy to control epidemic outbreaks. Connections between susceptible and infective individuals can be broken by either individual, and then reconnected to a randomly chosen member of the population. It is assumed that the reconnecting individual has no previous information on the epidemiological condition of the new contact. We show that reconnection can completely suppress the disease, both by continuous and discontinuous transitions between the endemic and the infection-free states. For diseases with an asymptomatic phase, we analyze the conditions for the suppression of the disease, and show that, even when these conditions are not met, the increase of the endemic infection level is usually rather small. We conclude that, within some simple epidemiological models, contact switching is a quite robust and effective control strategy. This suggests that it may also be an efficient method in more complex situations.
△ Less
Submitted 11 June, 2008;
originally announced June 2008.
-
The ground state energy of the Edwards-Anderson spin glass model with a parallel tempering Monte Carlo algorithm
Authors:
F. Roma,
S. Risau-Gusman,
A. J. Ramirez-Pastor,
F. Nieto,
E. E. Vogel
Abstract:
We study the efficiency of parallel tempering Monte Carlo technique for calculating true ground states of the Edwards-Anderson spin glass model. Bimodal and Gaussian bond distributions were considered in two and three-dimensional lattices. By a systematic analysis we find a simple formula to estimate the values of the parameters needed in the algorithm to find the GS with a fixed average probabi…
▽ More
We study the efficiency of parallel tempering Monte Carlo technique for calculating true ground states of the Edwards-Anderson spin glass model. Bimodal and Gaussian bond distributions were considered in two and three-dimensional lattices. By a systematic analysis we find a simple formula to estimate the values of the parameters needed in the algorithm to find the GS with a fixed average probability. We also study the performance of the algorithm for single samples, quantifying the difference between samples where the GS is hard, or easy, to find. The GS energies we obtain are in good agreement with the values found in the literature. Our results show that the performance of the parallel tempering technique is comparable to more powerful heuristics developed to find the ground state of Ising spin glass systems.
△ Less
Submitted 23 March, 2009; v1 submitted 5 June, 2008;
originally announced June 2008.
-
Fractal dimension of domain walls in the Edwards-Anderson spin glass model
Authors:
S. Risau-Gusman,
F. Roma
Abstract:
We study directly the length of the domain walls (DW) obtained by comparing the ground states of the Edwards-Anderson spin glass model subject to periodic and antiperiodic boundary conditions. For the bimodal and Gaussian bond distributions, we have isolated the DW and have calculated directly its fractal dimension $d_f$. Our results show that, even though in three dimensions $d_f$ is the same f…
▽ More
We study directly the length of the domain walls (DW) obtained by comparing the ground states of the Edwards-Anderson spin glass model subject to periodic and antiperiodic boundary conditions. For the bimodal and Gaussian bond distributions, we have isolated the DW and have calculated directly its fractal dimension $d_f$. Our results show that, even though in three dimensions $d_f$ is the same for both distributions of bonds, this is clearly not the case for two-dimensional (2D) systems. In addition, contrary to what happens in the case of the 2D Edwards-Anderson spin glass with Gaussian distribution of bonds, we find no evidence that the DW for the bimodal distribution of bonds can be described as a Schramm-Loewner evolution processes.
△ Less
Submitted 27 March, 2008; v1 submitted 1 November, 2007;
originally announced November 2007.
-
Oscillations in two-species models: tying the stochastic and deterministic approaches
Authors:
Sebastián Risau-Gusman,
Guillermo Abramson
Abstract:
We analyze general two-species stochastic models, of the kind generally used for the study of population dynamics. We show that the conditions for the stochastic (microscopic) model to display approximate sustained oscillatory behavior are governed by the parameters of the corresponding deterministic (macroscopic) model. We provide a quantitative criterion for the quality of the stochastic oscil…
▽ More
We analyze general two-species stochastic models, of the kind generally used for the study of population dynamics. We show that the conditions for the stochastic (microscopic) model to display approximate sustained oscillatory behavior are governed by the parameters of the corresponding deterministic (macroscopic) model. We provide a quantitative criterion for the quality of the stochastic oscillation, using a dimensionless parameter that depends only on the deterministic model. When this parameter is small, the oscillations are clear, and the frequencies of the stochastic and deterministic oscillations are close, for all stochastic models compatible with the same deterministic one. On the other hand, when it is large, the oscillations cannot be distinguished from a noise.
△ Less
Submitted 26 January, 2007;
originally announced January 2007.
-
Influence of the Ground-State Topology on the Domain-Wall Energy in the Edwards-Anderson +/- J Spin Glass Model
Authors:
F. Roma,
S. Risau-Gusman,
A. J. Ramirez-Pastor,
F. Nieto,
E. E. Vogel
Abstract:
We study the phase stability of the Edwards-Anderson spin-glass model by analyzing the domain-wall energy. For the bimodal distribution of bonds, a topological analysis of the ground state allows us to separate the system into two regions: the backbone and its environment. We find that the distributions of domain-wall energies are very different in these two regions for the three dimensional (3D…
▽ More
We study the phase stability of the Edwards-Anderson spin-glass model by analyzing the domain-wall energy. For the bimodal distribution of bonds, a topological analysis of the ground state allows us to separate the system into two regions: the backbone and its environment. We find that the distributions of domain-wall energies are very different in these two regions for the three dimensional (3D) case. Although the backbone turns out to have a very high phase stability, the combined effect of these excitations and correlations produces the low global stability displayed by the system as a whole. On the other hand, in two dimensions (2D) we find that the surface of the excitations avoids the backbone. Our results confirm that a narrow connection exists between the phase stability of the system and the internal structure of the ground-state. In addition, for both 3D and 2D we are able to obtain the fractal dimension of the domain wall by direct means.
△ Less
Submitted 21 December, 2006; v1 submitted 23 July, 2006;
originally announced July 2006.
-
Topology, phase transitions and the spherical model
Authors:
S. Risau-Gusman,
A. C. Ribeiro-Teixeira,
D. A. Stariolo
Abstract:
The Topological Hypothesis states that phase transitions should be related to changes in the topology of configuration space. The necessity of such changes has already been demonstrated. We characterize exactly the topology of the configuration space of the short range Berlin-Kac spherical model, for spins lying in hypercubic lattices of dimension d. We find a continuum of changes in the topolog…
▽ More
The Topological Hypothesis states that phase transitions should be related to changes in the topology of configuration space. The necessity of such changes has already been demonstrated. We characterize exactly the topology of the configuration space of the short range Berlin-Kac spherical model, for spins lying in hypercubic lattices of dimension d. We find a continuum of changes in the topology and also a finite number of discontinuities in some topological functions. We show however that these discontinuities do not coincide with the phase transitions which happen for d >= 3, and conversely, that no topological discontinuity can be associated to them. This is the first short range, confining potential for which the existence of special topological changes are shown not to be sufficient to infer the occurrence of a phase transition.
△ Less
Submitted 18 August, 2005;
originally announced August 2005.
-
Properties of dense partially random graphs
Authors:
Sebastian Risau-Gusman
Abstract:
We study the properties of random graphs where for each vertex a {\it neighbourhood} has been previously defined. The probability of an edge joining two vertices depends on whether the vertices are neighbours or not, as happens in Small World Graphs (SWGs). But we consider the case where the average degree of each node is of order of the size of the graph (unlike SWGs, which are sparse). This al…
▽ More
We study the properties of random graphs where for each vertex a {\it neighbourhood} has been previously defined. The probability of an edge joining two vertices depends on whether the vertices are neighbours or not, as happens in Small World Graphs (SWGs). But we consider the case where the average degree of each node is of order of the size of the graph (unlike SWGs, which are sparse). This allows us to calculate the mean distance and clustering, that are qualitatively similar (although not in such a dramatic scale range) to the case of SWGs. We also obtain analytically the distribution of eigenvalues of the corresponding adjacency matrices. This distribution is discrete for large eigenvalues and continuous for small eigenvalues. The continuous part of the distribution follows a semicircle law, whose width is proportional to the "disorder" of the graph, whereas the discrete part is simply a rescaling of the spectrum of the substrate. We apply our results to the calculation of the mixing rate and the synchronizability threshold.
△ Less
Submitted 28 September, 2004;
originally announced September 2004.
-
The dynamics of opinion in hierarchical organizations
Authors:
M. F. Laguna,
S. Risau-Gusman,
G. Abramson,
S. Goncalves,
J. R. Iglesias
Abstract:
We study the mutual influence of authority and persuasion in the flow of opinion. Many social organizations are characterized by a hierarchical structure where the propagation of opinion is asymmetric. In the normal flow of opinion formation a high-rank agent uses its authority (or its persuasion when necessary) to impose its opinion on others. However, agents with no authority may only use the…
▽ More
We study the mutual influence of authority and persuasion in the flow of opinion. Many social organizations are characterized by a hierarchical structure where the propagation of opinion is asymmetric. In the normal flow of opinion formation a high-rank agent uses its authority (or its persuasion when necessary) to impose its opinion on others. However, agents with no authority may only use the force of its persuasion to propagate their opinions. In this contribution we describe a simple model with no social mobility, where each agent belongs to a class in the hierarchy and has also a persuasion capability. The model is studied numerically for a three levels case, and analytically within a mean field approximation, with a very good agreement between the two approaches. The stratum where the dominant opinion arises from is strongly dependent on the percentage of agents in each hierarchy level, and we obtain a phase diagram identifying the relative frequency of prevailing opinions. We also find that the time evolution of the conflicting opinions polarizes after a short transient.
△ Less
Submitted 12 April, 2004;
originally announced April 2004.
-
Esca** from cycles through a glass transition
Authors:
Sebastian Risau-Gusman,
Alexandre S. Martinez,
Osame Kinouchi
Abstract:
A random walk is performed over a disordered media composed of $N$ sites random and uniformly distributed inside a $d$-dimensional hypercube. The walker cannot remain in the same site and hops to one of its $n$ neighboring sites with a transition probability that depends on the distance $D$ between sites according to a cost function $E(D)$. The stochasticity level is parametrized by a formal tem…
▽ More
A random walk is performed over a disordered media composed of $N$ sites random and uniformly distributed inside a $d$-dimensional hypercube. The walker cannot remain in the same site and hops to one of its $n$ neighboring sites with a transition probability that depends on the distance $D$ between sites according to a cost function $E(D)$. The stochasticity level is parametrized by a formal temperature $T$. In the case $T = 0$, the walk is deterministic and ergodicity is broken: the phase space is divided in a ${\cal O}(N)$ number of attractor basins of two-cycles that trap the walker. For $d = 1$, analytic results indicate the existence of a glass transition at $T_1 = 1/2$ as $N \to \infty$. Below $T_1$, the average trap** time in two-cycles diverges and out-of-equilibrium behavior appears. Similar glass transitions occur in higher dimensions choosing a proper cost function. We also present some results for the statistics of distances for Poisson spatial point processes.
△ Less
Submitted 13 March, 2003; v1 submitted 10 January, 2003;
originally announced January 2003.
-
Exploratory Behavior, Trap Models and Glass Transitions
Authors:
Alexandre S. Martinez,
Osame Kinouchi,
Sebastian Risau-Gusman
Abstract:
A random walk is performed on a disordered landscape composed of $N$ sites randomly and uniformly distributed inside a $d$-dimensional hypercube. The walker hops from one site to another with probability proportional to $\exp [- βE(D)]$, where $β= 1/T$ is the inverse of a formal temperature and $E(D)$ is an arbitrary cost function which depends on the hop distance $D$. Analytic results indicate…
▽ More
A random walk is performed on a disordered landscape composed of $N$ sites randomly and uniformly distributed inside a $d$-dimensional hypercube. The walker hops from one site to another with probability proportional to $\exp [- βE(D)]$, where $β= 1/T$ is the inverse of a formal temperature and $E(D)$ is an arbitrary cost function which depends on the hop distance $D$. Analytic results indicate that, if $E(D) = D^{d}$ and $N \to \infty$, there exists a glass transition at $β_d = π^{d/2}/Γ(d/2 + 1)$. Below $T_d$, the average trap** time diverges and the system falls into an out-of-equilibrium regime with aging phenomena. A Lévy flight scenario and applications to exploratory behavior are considered.
△ Less
Submitted 13 June, 2003; v1 submitted 24 October, 2002;
originally announced October 2002.
-
Learning curves for Soft Margin Classifiers
Authors:
Sebastian Risau-Gusman,
Mirta B. Gordon
Abstract:
Typical learning curves for Soft Margin Classifiers (SMCs) learning both realizable and unrealizable tasks are determined using the tools of Statistical Mechanics. We derive the analytical behaviour of the learning curves in the regimes of small and large training sets. The generalization errors present different decay laws towards the asymptotic values as a function of the training set size, de…
▽ More
Typical learning curves for Soft Margin Classifiers (SMCs) learning both realizable and unrealizable tasks are determined using the tools of Statistical Mechanics. We derive the analytical behaviour of the learning curves in the regimes of small and large training sets. The generalization errors present different decay laws towards the asymptotic values as a function of the training set size, depending on general geometrical characteristics of the rule to be learned. Optimal generalization curves are deduced through a fine tuning of the hyperparameter controlling the trade-off between the error and the regularization terms in the cost function. Even if the task is realizable, the optimal performance of the SMC is better than that of a hard margin Support Vector Machine (SVM) learning the same rule, and is very close to that of the Bayesian classifier.
△ Less
Submitted 14 March, 2002;
originally announced March 2002.
-
Deterministic walks in random networks: an application to thesaurus graphs
Authors:
O. Kinouchi,
A. S. Martinez,
G. F. Lima,
G. M. Lourenco,
S. Risau-Gusman
Abstract:
In a landscape composed of N randomly distributed sites in Euclidean space, a walker (``tourist'') goes to the nearest one that has not been visited in the last τsteps. This procedure leads to trajectories composed of a transient part and a final cyclic attractor of period p. The tourist walk presents universal aspects with respect to τand can be done in a wide range of networks that can be view…
▽ More
In a landscape composed of N randomly distributed sites in Euclidean space, a walker (``tourist'') goes to the nearest one that has not been visited in the last τsteps. This procedure leads to trajectories composed of a transient part and a final cyclic attractor of period p. The tourist walk presents universal aspects with respect to τand can be done in a wide range of networks that can be viewed as ordinal neighborhood graphs. As an example, we show that graphs defined by thesaurus dictionaries share some of the statistical properties of low dimensional (d=2) Euclidean graphs and are easily distinguished from random graphs. This approach furnishes complementary information to the usual clustering coefficient and mean minimum separation length.
△ Less
Submitted 23 April, 2002; v1 submitted 10 October, 2001;
originally announced October 2001.
-
Statistical Mechanics of Soft Margin Classifiers
Authors:
Sebastian Risau-Gusman,
Mirta B. Gordon
Abstract:
We study the typical learning properties of the recently introduced Soft Margin Classifiers (SMCs), learning realizable and unrealizable tasks, with the tools of Statistical Mechanics. We derive analytically the behaviour of the learning curves in the regime of very large training sets. We obtain exponential and power laws for the decay of the generalization error towards the asymptotic value, d…
▽ More
We study the typical learning properties of the recently introduced Soft Margin Classifiers (SMCs), learning realizable and unrealizable tasks, with the tools of Statistical Mechanics. We derive analytically the behaviour of the learning curves in the regime of very large training sets. We obtain exponential and power laws for the decay of the generalization error towards the asymptotic value, depending on the task and on general characteristics of the distribution of stabilities of the patterns to be learned. The optimal learning curves of the SMCs, which give the minimal generalization error, are obtained by tuning the coefficient controlling the trade-off between the error and the regularization terms in the cost function. If the task is realizable by the SMC, the optimal performance is better than that of a hard margin Support Vector Machine and is very close to that of a Bayesian classifier.
△ Less
Submitted 18 February, 2001;
originally announced February 2001.
-
Hierarchical learning in polynomial Support Vector Machines
Authors:
Sebastian Risau-Gusman,
Mirta B. Gordon
Abstract:
We study the typical properties of polynomial Support Vector Machines within a Statistical Mechanics approach that allows us to analyze the effect of different normalizations of the features. If the normalization is adecuately chosen, there is a hierarchical learning of features of increasing order as a function of the training set size.
We study the typical properties of polynomial Support Vector Machines within a Statistical Mechanics approach that allows us to analyze the effect of different normalizations of the features. If the normalization is adecuately chosen, there is a hierarchical learning of features of increasing order as a function of the training set size.
△ Less
Submitted 26 October, 2000;
originally announced October 2000.
-
Generalization properties of finite size polynomial Support Vector Machines
Authors:
Sebastian Risau-Gusman,
Mirta B. Gordon
Abstract:
The learning properties of finite size polynomial Support Vector Machines are analyzed in the case of realizable classification tasks. The normalization of the high order features acts as a squeezing factor, introducing a strong anisotropy in the patterns distribution in feature space. As a function of the training set size, the corresponding generalization error presents a crossover, more or le…
▽ More
The learning properties of finite size polynomial Support Vector Machines are analyzed in the case of realizable classification tasks. The normalization of the high order features acts as a squeezing factor, introducing a strong anisotropy in the patterns distribution in feature space. As a function of the training set size, the corresponding generalization error presents a crossover, more or less abrupt depending on the distribution's anisotropy and on the task to be learned, between a fast-decreasing and a slowly decreasing regime. This behaviour corresponds to the stepwise decrease found by Dietrich et al.[Phys. Rev. Lett. 82 (1999) 2975-2978] in the thermodynamic limit. The theoretical results are in excellent agreement with the numerical simulations.
△ Less
Submitted 4 February, 2000;
originally announced February 2000.
