International Journal of Neural Systems, Jun 1, 1994
We study the storage properties associated with generalized Hebbian learning rules which present ... more We study the storage properties associated with generalized Hebbian learning rules which present four free parameters that allow for asymmetry. We also introduce two extra parameters in the post-synaptic potentials in order to further improve the critical capacity. Using signal-to-noise analysis, as well as computer simulations on an analog network, we discuss the performance of the rules for arbitrarily biased patterns and find that the critical storage capacity alpha c becomes maximal for a particular symmetric rule (alpha c diverges in the sparse coding limit). Departures from symmetry decrease alpha c but can increase the robustness of the model.
Particle motion of a Lennard-Jones supercooled liquid near the glass transition is studied by mol... more Particle motion of a Lennard-Jones supercooled liquid near the glass transition is studied by molecular dynamics simulations. The van Hove distribution of particle displacements and the self part of the incoherent scattering function on a large window of time and length scales are measured in order to characterize particle diffusion. We consider inherent structures dynamics in order to filter out vibrational motion. We find at least three clear regimes in both functions: a persistence dominated regime at short times and small length scales, a strongly correlated regime of particle motion at intermediate time-lengths and a Gaussian or Fickian asymptotic regime at long times/lengths. The crossovers between these regimes allow us to obtain two characteristic length scales from two point functions. The functional forms of the observables and time/length scalings are considered. A comparison with available experimental results on local particle motion is done.
ABSTRACT We study the phase structure of Ising spin glass models with interactions between 3, 5 a... more ABSTRACT We study the phase structure of Ising spin glass models with interactions between 3, 5 and 10 spins in the Hamiltonian. We used the replica mean field theory approach and solved numerically the saddle point equations. We found evidence of three phases: a disordered phase and two spin glass phases at low temperatures. Present address: Centro Brasileiro de Pesquisas Físicas (CBPF), Rua Xavier Sigaud 150, 22290 Rio de Janeiro, RJ, Brazil.
International Journal of Neural Systems, Jun 1, 1994
We study the storage properties associated with generalized Hebbian learning rules which present ... more We study the storage properties associated with generalized Hebbian learning rules which present four free parameters that allow for asymmetry. We also introduce two extra parameters in the post-synaptic potentials in order to further improve the critical capacity. Using signal-to-noise analysis, as well as computer simulations on an analog network, we discuss the performance of the rules for arbitrarily biased patterns and find that the critical storage capacity alpha c becomes maximal for a particular symmetric rule (alpha c diverges in the sparse coding limit). Departures from symmetry decrease alpha c but can increase the robustness of the model.
Particle motion of a Lennard-Jones supercooled liquid near the glass transition is studied by mol... more Particle motion of a Lennard-Jones supercooled liquid near the glass transition is studied by molecular dynamics simulations. The van Hove distribution of particle displacements and the self part of the incoherent scattering function on a large window of time and length scales are measured in order to characterize particle diffusion. We consider inherent structures dynamics in order to filter out vibrational motion. We find at least three clear regimes in both functions: a persistence dominated regime at short times and small length scales, a strongly correlated regime of particle motion at intermediate time-lengths and a Gaussian or Fickian asymptotic regime at long times/lengths. The crossovers between these regimes allow us to obtain two characteristic length scales from two point functions. The functional forms of the observables and time/length scalings are considered. A comparison with available experimental results on local particle motion is done.
ABSTRACT We study the phase structure of Ising spin glass models with interactions between 3, 5 a... more ABSTRACT We study the phase structure of Ising spin glass models with interactions between 3, 5 and 10 spins in the Hamiltonian. We used the replica mean field theory approach and solved numerically the saddle point equations. We found evidence of three phases: a disordered phase and two spin glass phases at low temperatures. Present address: Centro Brasileiro de Pesquisas Físicas (CBPF), Rua Xavier Sigaud 150, 22290 Rio de Janeiro, RJ, Brazil.
Uploads
Papers by Daniel Stariolo