{
  "id": "cond-mat/0305587",
  "version": "v1",
  "published": "2003-05-26T08:30:33.000Z",
  "updated": "2003-05-26T08:30:33.000Z",
  "title": "A layered neural network with three-state neurons optimizing the mutual information",
  "authors": [
    "D. Bolle",
    "R. Erichsen, Jr.",
    "W. K. Theumann"
  ],
  "comment": "17 pages Latex including 6 eps-figures",
  "doi": "10.1016/j.physa.2003.10.033",
  "categories": [
    "cond-mat.dis-nn",
    "cond-mat.stat-mech",
    "q-bio"
  ],
  "abstract": "The time evolution of an exactly solvable layered feedforward neural network with three-state neurons and optimizing the mutual information is studied for arbitrary synaptic noise (temperature). Detailed stationary temperature-capacity and capacity-activity phase diagrams are obtained. The model exhibits pattern retrieval, pattern-fluctuation retrieval and spin-glass phases. It is found that there is an improved performance in the form of both a larger critical capacity and information content compared with three-state Ising-type layered network models. Flow diagrams reveal that saddle-point solutions associated with fluctuation overlaps slow down considerably the flow of the network states towards the stable fixed-points.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2003-05-26T08:30:33.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "layered neural network",
      "mutual information",
      "three-state neurons optimizing",
      "layered feedforward neural network",
      "solvable layered feedforward neural"
    ],
    "tags": [
      "journal article"
    ],
    "note": {
      "typesetting": "LaTeX",
      "pages": 17,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}