Searcharxiv⌕ Search

SEARCH · Searcharxiv

Results for “adap-org”

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 379 records · Page 21Linked to original sources
arXiv1995

Strong Ordering by Non-uniformity of Thresholds in a Coupled Map Lattice

The coupled map lattice by Olami {\it et al.} [Phys. Rev. Lett. {\bf 68}, 1244 (1992)] is ``doped'' by letting just {\it one} site have a threshold, $T^{*}_{\rm max}$, bigger than the others. On an $L \times L$ lattice with periodic boundary conditions this leads to a transition from avalanche sizes of about one to exactly $L^2$, and after each avalanche stresses distributes among only five distinct values, $τ_{k}$, related to the parameters $α$ and $T^{*}_{\rm max}$ by $τ_{k}= kα T^{*}_{\rm max}$ where $k=0,1,2,3,4$. This result is independent of lattice size. The transient times are inversely proportional to the amount of doping and increase linearly with $L$.

adap-org↗
arXiv1993

Exact solutions for a mean-field Abelian sandpile

We introduce a model for a sandpile, with N sites, critical height N and each site connected to every other site. It is thus a mean-field model in the spin-glass sense. We find an exact solution for the steady state probability distribution of avalanche sizes, and discuss its asymptotics for large N.

adap-org↗
arXiv1996

The Dynamics of a Genetic Algorithm for a Simple Learning Problem

A formalism for describing the dynamics of Genetic Algorithms (GAs) using methods from statistical mechanics is applied to the problem of generalization in a perceptron with binary weights. The dynamics are solved for the case where a new batch of training patterns is presented to each population member each generation, which considerably simplifies the calculation. The theory is shown to agree closely to simulations of a real GA averaged over many runs, accurately predicting the mean best solution found. For weak selection and large problem size the difference equations describing the dynamics can be expressed analytically and we find that the effects of noise due to the finite size of each training batch can be removed by increasing the population size appropriately. If this population resizing is used, one can deduce the most computationally efficient size of training batch each generation. For independent patterns this choice also gives the minimum total number of training patterns used. Although using independent patterns is a very inefficient use of training patterns in general, this work may also prove useful for determining the optimum batch size in the case where patterns are recycled.

cond-mat↗
arXiv1995

Phase space geometry and slow dynamics

We describe a non-Arrhenius mechanism for slowing down of dynamics that is inherent to the high dimensionality of the phase space. We show that such a mechanism is at work both in a family of mean-field spin-glass models without any domain structure and in the case of ferromagnetic domain growth. The marginality of spin-glass dynamics, as well as the existence of a `quasi equilibrium regime' can be understood within this scenario. We discuss the question of ergodicity in an out-of equilibrium situation.

cond-mat↗
arXiv1997

Car accidents and number of stopped cars due to road blockage on a one-lane highway

Within the framework of a simple model of car traffic on a one-lane highway, we study the probability for car accidents to occur when drivers do not respect the safety distance between cars, and, as a result of the blockage during the time $T$ necessary to clear the road, we determine the number of stopped cars as a function of car density. We give a simple theory in good agreement with our numerical simulations.

adap-org↗
arXiv1997

On-line learning of non-monotonic rules by simple perceptron

We study the generalization ability of a simple perceptron which learns unlearnable rules. The rules are presented by a teacher perceptron with a non-monotonic transfer function. The student is trained in the on-line mode. The asymptotic behaviour of the generalization error is estimated under various conditions. Several learning strategies are proposed and improved to obtain the theoretical lower bound of the generalization error.

cond-mat↗
arXiv1996

Error threshold in simple landscapes

We consider the quasispecies description of a population evolving in both the "master sequence" landscape (where a single sequence is evolutionarily preferred over all others) and the REM landscape (where the fitness of different sequences is an independent, identically distributed, random variable). We show that, in both cases, the error threshold is analogous to a first order thermodynamical transition, where the overlap between the average genotype and the optimal one drops discontinuously to zero.

cond-mat.dis-nn↗
arXiv1998

Entropic Sampling and Natural Selection in Biological Evolution

With a view to connecting random mutation on the molecular level to punctuated equilibrium behavior on the phenotype level, we propose a new model for biological evolution, which incorporates random mutation and natural selection. In this scheme the system evolves continuously into new configurations, yielding non-stationary behavior of the total fitness. Further, both the waiting time distribution of species and the avalanche size distribution display power-law behaviors with exponents close to two, which are consistent with the fossil data. These features are rather robust, indicating the key role of entropy.

cond-mat↗
arXiv1996

Error threshold in the evolution of diploid organisms

The effects of error propagation in the reproduction of diploid organisms are studied within the populational genetics framework of the quasispecies model. The dependence of the error threshold on the dominance parameter is fully investigated. In particular, it is shown that dominance can protect the wild-type alleles from the error catastrophe. The analysis is restricted to a diploid analogue of the single-peaked landscape.

cond-mat.dis-nn↗
arXiv1998

Probabilistic Analysis of the Number Partitioning Problem

Given a sequence of $N$ positive real numbers $\{a_1,a_2,..., a_N \}$, the number partitioning problem consists of partitioning them into two sets such that the absolute value of the difference of the sums of $a_j$ over the two sets is minimized. In the case that the $a_j$'s are statistically independent random variables uniformly distributed in the unit interval, this NP-complete problem is equivalent to the problem of finding the ground state of an infinite-range, random anti-ferromagnetic Ising model. We employ the annealed approximation to derive analytical lower bounds to the average value of the difference for the best constrained and unconstrained partitions in the large $N$ limit. Furthermore, we calculate analytically the fraction of metastable states, i.e. states that are stable against all single spin flips, and found that it vanishes like $N^{-3/2}$.

adap-org↗
arXiv1998

Cellular automaton rules conserving the number of active sites

This paper shows how to determine all the unidimensional two-state cellular automaton rules of a given number of inputs which conserve the number of active sites. These rules have to satisfy a necessary and sufficient condition. If the active sites are viewed as cells occupied by identical particles, these cellular automaton rules represent evolution operators of systems of identical interacting particles whose total number is conserved. Some of these rules, which allow motion in both directions, mimic ensembles of one-dimensional pseudo-random walkers. Numerical evidence indicates that the corresponding stochastic processes might be non-Gaussian.

adap-org↗
arXiv1997

Noise induced transport at zero temperature

We consider a particle in the over-damped regime at zero temperature under the influence of a sawtooth potential and of a noisy force, which is correlated in time. A current occurs, even if the mean of the noisy force vanishes. We calculate the stationary probability distribution and the stationary current. We discuss, how these items depend on the characteristic parameters of the underlying stochastic process. A formal expansion of the current around the white-noise limit not always gives the correct asymptotic behaviour. We improve the expansion for some simple but representative cases.

cond-mat.stat-mech↗
arXiv1999

Extremal Optimization of Graph Partitioning at the Percolation Threshold

The benefits of a recently proposed method to approximate hard optimization problems are demonstrated on the graph partitioning problem. The performance of this new method, called Extremal Optimization, is compared to Simulated Annealing in extensive numerical simulations. While generally a complex (NP-hard) problem, the optimization of the graph partitions is particularly difficult for sparse graphs with average connectivities near the percolation threshold. At this threshold, the relative error of Simulated Annealing for large graphs is found to diverge relative to Extremal Optimization at equalized runtime. On the other hand, Extremal Optimization, based on the extremal dynamics of self-organized critical systems, reproduces known results about optimal partitions at this critical point quite well.

cond-mat.stat-mech↗
arXiv1999

Enhanced winnings in a mixed-ability population playing a minority game

We study a mixed population of adaptive agents with small and large memories, competing in a minority game. If the agents are sufficiently adaptive, we find that the average winnings per agent can exceed that obtainable in the corresponding pure populations. In contrast to the pure population, the average success rate of the large-memory agents can be greater than 50 percent. The present results are not reproduced if the agents are fed a random history, thereby demonstrating the importance of memory in this system.

cond-mat↗
arXiv2000

Critical behavior of a cellular automaton highway traffic model

We derive the critical behavior of a CA traffic flow model using an order parameter breaking the symmetry of the jam-free phase. Random braking appears to be the symmetry-breaking field conjugate to the order parameter. For $v_{\max}=2$, we determine the values of the critical exponents $β$, $γ$ and $δ$ using an order-3 cluster approximation and computer simulations. These critical exponents satisfy a scaling relation, which can be derived assuming that the order parameter is a generalized homogeneous function of $|ρ-ρ_c|$ and p in the vicinity of the phase transition point.

cond-mat.dis-nn↗
arXiv1997

Polarization fluctuations in vertical cavity surface emitting lasers: a key to the mechanism behind polarization stability

We investigate the effects of the electron-hole spin dynamics on the polarization fluctuations in the light emitted from a vertical cavity surface emitting laser (VCSEL). The Langevin equations are derived based on a rate equation model including birefringence, dichroism, and two carrier density pools seperately coupled to right and left circular polarization. The results show that the carrier dynamics phase lock the polarization fluctuations to the laser mode. This is clearly seen in the difference between fluctuations in ellipticity and fluctuations in polarization direction. Seperate measurements of the polarization fluctuations in ellipticity and in polarization direction can therefore provide quantitative information on the non-linear contribution of the carrier dynamics to polarization stability in VCSELs.

physics.optics↗
arXiv1997

The split density model: a unified description of polarization and array dynamics for vertical cavity surface emitting lasers

A rate equation model describing the laser dynamics of general one or two dimensional vertical cavity surface emitting laser (vcsel) arrays is introduced. It is shown that the theory includes both the previous theory for edge emitting semiconductor laser arrays and the theory of polarization dynamics in single quantum well vcsels in a single unified description. The model is based on the physical assumption of separate carrier density pools individually coupled to different light field modes. These modes interact through the coherent light field dynamics derived from Maxwell's equations. The special case of two densities and two light field modes is solved and the implications for larger arrays are discussed. Our analytic results show that typical solutions of the split density model range from phase locking to chaos, depending on the magnitude of the coherent interaction. For weak coupling the stable supermode is always the mode of highest frequency. This indicates that anti-phase locking is the only stable phase locking possible in semiconductor laser arrays.

physics.optics↗