Presentation | 2004/1/20 The Frustrated Lattice Gas Model and Associative Memories (Neurocomputing) Jun-ichi INOUE, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Associative memories are formulated by the frustrated lattice gas (FLG) model. In this mathematical model, each neuron is labeled as "hole" or "non-hole" (namely, "occupied") dynamically. Consequently, the hole neuron is isolated from the neural network. The number of the hole neurons is controlled by chemical potential of the system. With the assistance of the replica method, we investigate the performance of the dynamically diluted network as associative memories. We find that in comparison with the conventional Hopfield model, the pattern-retrieval ability of this network is easily weaken by the thermal noise, however, the network is robust against the increase of the cross-talk noise. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | associative memories / frustrated lattice gas / statistical mechanics / replica method / phase transition |
Paper # | NC2003-130 |
Date of Issue |
Conference Information | |
Committee | NC |
---|---|
Conference Date | 2004/1/20(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Neurocomputing (NC) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | The Frustrated Lattice Gas Model and Associative Memories (Neurocomputing) |
Sub Title (in English) | |
Keyword(1) | associative memories |
Keyword(2) | frustrated lattice gas |
Keyword(3) | statistical mechanics |
Keyword(4) | replica method |
Keyword(5) | phase transition |
1st Author's Name | Jun-ichi INOUE |
1st Author's Affiliation | Graduate School of Enginnering, Hokkaido University() |
Date | 2004/1/20 |
Paper # | NC2003-130 |
Volume (vol) | vol.103 |
Number (no) | 602 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |