International Journal of Modern Physics C

We are fascinated by the idea of giving life to the inanimate. The fields of Artificial Life and Artificial Intelligence (AI) attempt to use a scientific approach to pursue this desire. The first steps on this approach hark back to Turing and his suggestion of an imitation game as an alternative answer to the question "can machines think?".¹ To test his hypothesis, Turing formulated the Turing test¹ to detect human behavior in computers. But how do humans pass such a test? What would you say if you would learn that they do not pass it well? What would it mean for our understanding of human behavior? What would it mean for our design of tests of the success of artificial life? We report below an experiment in which men consistently failed the Turing test.

In lattice Boltzmann simulations particle groups — represented by scalar velocity distributions — are moved on a finite lattice. The size of these particle groups is not well-defined although it is crucial to assume that they should be big enough for using a continuous distribution. Here we propose to use the liquid–vapor interface as an internal yardstick to scale the system. Comparison with existing experimental data and with molecular dynamics simulation of Lennard–Jones-argon shows that the number of atoms located on one lattice site is in the order of few atoms. This contradicts the initial assumption concerning the number of particles in the group, therefore seems to raise some doubts about the applicability of the lattice Boltzmann method in certain problems whenever interfaces play important role and ergodicity does not hold.

Cellular automata (CA) are discrete dynamical systems that can be employed as cryptography algorithms. Irreversible CA with toggle rules — either to the right or to the left — were already proposed for encrypting. However, the similarity between ciphertexts when the plaintext is submitted to a small perturbation was identified as a flaw in this model. Here, such a model is altered by using bi-directional toggle CA rules instead of one-directional ones. Numerical experiments showed that the similarity flaw was eliminated by considering this modification, obtaining protection against differential cryptanalysis.

Using the new supercomputer JUMP at the Research Center Jülich, we were able to simulate large lattices (up to L=2·10⁶, a new world record) for long times (up to T=6000 for L=1.5·10⁵). Using this data, we examined the dynamical critical exponent z. The old assumption of z=2 with logarithmic corrections seems very unlikely according to our data, leaving the asymptotic value of z≃2.167.

We study the discrete Gierer–Meinhardt model of reaction–diffusion on three different types of networks: regular, random and scale-free. The model dynamics lead to the formation of stationary Turing patterns in the steady state in certain parameter regions. Some general features of the patterns are studied through numerical simulation. The results for the random and scale-free networks show a marked difference from those in the case of the regular network. The difference may be ascribed to the small world character of the first two types of networks.

The usual two-body score (energy) function to recognize native folds of proteins is Miyazawa–Jernigan (MJ) pairwise-contact function. The pairwise-contact parameters between two amino acids in MJ function are symmetric in a sense that a directional order of amino acids sequence along the backbone of a protein is ignored in constructing score parameters. Here we report that we succeeded in constructing a nonsymmetric two-body score function, capturing a directional order of amino acids sequence, by a perceptron learning and a protein threading. We considered pairs of two adjacent amino acids that are separated by two consecutive peptide bonds with the backbone directionality from the N-terminus to the C-terminus of a protein. We also considered the local environmental character, such as the secondary structures and the hydrophobicity (solvation), of amino acids in protein structures. The score is a corresponding propensity for a directional alignment of these two adjacent amino acids with their local environments. The resulting score function simultaneously recognized native folds of 1006 proteins covering all representative proteins with a homology less than 30% among them. The quality of this score function was validated by a threading test of new distinct 382 proteins with a homology less than 90% among them, and it entailed a high success ratio for recognizing native folds of 364 (95.3%) proteins. It showed a good feasibility of designing protein score functions for protein fold recognition by a perceptron learning and a protein threading.

We study ground state pairwise entanglement within one-dimensional spin-1/2 antiferromagnetic J₁–J₂ model with competing interactions. Contrary to some claims we found that frustration does not increase entanglement. Concurrence of nearest and next nearest neighbors are found to show abrupt change at phase transition points. We also show that the concurrence can be used to classify the phase diagram of the model in anisotropy–frustration plane.

A method is proposed to find the wave function of an electron moving infinitely in the field of an arbitrary 1D layered structure with two different homogeneous semi-infinite boundaries. It is shown that in general the problem reduces to solution of a set of two linear recurrent equations. The proposed approach is discussed on a base of two cases: a structure of periodically placed identical rectangular potentials and a nonordered structure with certain distortion of periodicity and potential identity.

A generalized magnetically controlled ballistic rain-like deposition (MBD) model of granular piles has been numerically investigated in 2D. The grains are taken to be elongated disks whence characterized by a two-state scalar degree of freedom, called "nip", their interaction being described through a Hamiltonian. The results are discussed in order to search for the effect of nip flip (or grain rotation from vertical to horizontal and conversely) probability in building a granular pile. The characteristics of creation of + (or -) nip's clusters and clusters of holes (missing nips) are analyzed. Two different cluster-mass regimes have been identified, through the cluster-mass distribution function which can be exponential or have a power law form depending on whether the nip flip (or grain rotation) probability is large or small. Analytical forms of the exponent are empirically found in terms of the Hamiltonian parameters.

The Lie series technique is applied to solve a class of nonlinear differential equations. We show that it includes also the dynamical integration technique. In particular systems with limit cycle behavior are studied. Differential equations with chaotic behavior are also considered with this technique.

One-dimensional cellular automata are dynamical systems characterized by discreteness (in space and time), determinism and local interaction. We present a procedure to calculate the ancestors for a given sequence of states, which is based on a special kind of graph called subset diagram. We use this diagram to specify subset tables for calculating ancestors which are not Garden-of-Eden sequences, hence the process is able to yield ancestors in several generations. Some examples are illustrated using the cellular automaton Rule 110 which is the most interesting automaton of two states and three neighbors.

The development of an efficient image-based computer identification system for plants or other organisms is an important ambitious goal, which is still far from realization. This paper presents three new methods potentially usable for such a system: fractal-based measures of complexity of leaf outline, a heuristic algorithm for automatic detection of leaf parts — the blade and the petiole, and a hierarchical perceptron — a kind of neural network classifier. The next few sets of automatically extractable features of leaf blades, encompassed those presented and/or traditionally used, are compared in the task of plant identification using the simplest known "nearest neighbor" identification algorithm, and more realistic neural network classifiers, especially the hierarchical. We show on two real data sets that the presented techniques are really usable for automatic identification, and are worthy of further investigation.