Boundary conditions and phase transitions in neural networks. Simulation results

This paper gives new simulation results on the asymptotic behaviour of theoretical neural networks on Z and Z(2) following an extended Hopfield law. It specifically focuses on the influence of fixed boundary conditions on such networks. First, we will generalise the theoretical results already obtained for attractive networks in one dimension to more complicated neural networks. Then, we will focus on two-dimensional neural networks. Theoretical results have already been found for the nearest neighbours Ising model in 2D with translation-invariant local isotropic interactions. We will detail what happens for this kind of interaction in neural networks and we will also focus on more complicated interactions, i.e., interactions that are not local, neither isotropic, nor translation-invariant. For all these kinds of interactions, we will show that fixed boundary conditions have significant impacts on the asymptotic behaviour of such networks. These impacts result in the emergence of phase transitions whose geometric shape will be numerically characterised.

[1]  A. Barabasi,et al.  Network biology: understanding the cell's functional organization , 2004, Nature Reviews Genetics.

[2]  J. Propp,et al.  Exact sampling with coupled Markov chains and applications to statistical mechanics , 1996 .

[3]  R. L. Dobrushin.  Existence of phase transitions in models of a lattice gas , 1967 .

[4]  Hiroaki Kitano,et al.  Biological robustness , 2008, Nature Reviews Genetics.

[5]  J. Montoya,et al.  Small world patterns in food webs. , 2002, Journal of theoretical biology.

[6]  Albert-László Barabási,et al.  Statistical mechanics of complex networks , 2001, ArXiv.

[7]  J. Demongeot,et al.  Positive and negative feedback: striking a balance between necessary antagonists. , 2002, Journal of theoretical biology.

[8]  David Ruelle,et al.  On the use of “small external fields” in the problem of symmetry breakdown in statistical mechanics☆ , 1972 .

[9]  D. Ruelle.  Statistical Mechanics: Rigorous Results , 1999 .

[10]  Jacques Demongeot,et al.  High-dimensional Switches and the Modeling of Cellular Differentiation 2.2 Mathematical Models , 2022 .

[11]  Jacques Demongeot,et al.  Boundary conditions and phase transitions in neural networks. Theoretical results , 2008, Neural Networks.

[12]  E. Ising.  Beitrag zur Theorie des Ferromagnetismus , 1925 .

[13]  Martin Suter,et al.  Small World , 2002 .

[14]  Nazim Fatès,et al.  Cellular Automata , 2004, Lecture Notes in Computer Science.

[15]  L. Onsager.  Crystal statistics. I. A two-dimensional model with an order-disorder transition , 1944 .

[16]  H. Hinrichsen.  Non-equilibrium critical phenomena and phase transitions into absorbing states , 2000, cond-mat/0001070.

[17]  R. Dobrushin.  The problem of uniqueness of a gibbsian random field and the problem of phase transitions , 1968 .

[18]  Geoffrey E. Hinton,et al.  A Learning Algorithm for Boltzmann Machines , 1985, Cogn. Sci..

[19]  Duncan J. Watts,et al.  Collective dynamics of ‘small-world’ networks , 1998, Nature.

[20]  Nazim Fatès,et al.  An Experimental Study of Robustness to Asynchronism for Elementary Cellular Automata , 2004, Complex Syst..

[21]  Frank L. Spitzer,et al.  Introduction Aux Processus De Markov A Parametre Dans Zν , 1974 .

[22]  Hugues Berry,et al.  Nonequilibrium phase transition in a self-activated biological network. , 2003, Physical review. E, Statistical, nonlinear, and soft matter physics.

[23]  David Ruelle,et al.  Thermodynamic Formalism: The Mathematical Structure of Equilibrium Statistical Mechanics , 2004 .

[24]  Chandan Dasgupta,et al.  On the problem of spurious patterns in neural associative memory models , 1997, IEEE Trans. Neural Networks.