Neural node network and model, and method of teaching same


Ontology type: sgo:Patent     


Patent Info

DATE

1995-12-26T00:00

AUTHORS

Alexander G. Parlos , Amir F. Atiya , Benito Fernandez , Wei K. Tsai , Kil T. Chong

ABSTRACT

The present invention is a fully connected feed forward network that includes at least one hidden layer 16. The hidden layer 16 includes nodes 20 in which the output of the node is fed back to that node as an input with a unit delay produced by a delay device 24 occurring in the feedback path 22 (local feedback). Each node within each layer also receives a delayed output (crosstalk) produced by a delay unit 36 from all the other nodes within the same layer 16. The node performs a transfer function operation based on the inputs from the previous layer and the delayed outputs. The network can be implemented as analog or digital or within a general purpose processor. Two teaching methods can be used: (1) back propagation of weight calculation that includes the local feedback and the crosstalk or (2) more preferably a feed forward gradient decent which immediately follows the output computations and which also includes the local feedback and the crosstalk. Subsequent to the gradient propagation, the weights can be normalized, thereby preventing convergence to a local optimum. Education of the network can be incremental both on and off-line. An educated network is suitable for modeling and controlling dynamic nonlinear systems and time series systems and predicting the outputs as well as hidden states and parameters. The educated network can also be further educated during on-line processing. More... »

Related SciGraph Publications

  • 1989-12. Approximation by superpositions of a sigmoidal function in MATHEMATICS OF CONTROL, SIGNALS, AND SYSTEMS
  • JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/2746", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "name": "Alexander G. Parlos", 
            "type": "Person"
          }, 
          {
            "name": "Amir F. Atiya", 
            "type": "Person"
          }, 
          {
            "name": "Benito Fernandez", 
            "type": "Person"
          }, 
          {
            "name": "Wei K. Tsai", 
            "type": "Person"
          }, 
          {
            "name": "Kil T. Chong", 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "https://doi.org/10.1016/0005-1098(90)90154-a", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1001974240"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/0005-1098(90)90154-a", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1001974240"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1162/neco.1990.2.4.490", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1003959523"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1162/neco.1989.1.2.161", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1007499213"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/bf02551274", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023250347", 
              "https://doi.org/10.1007/bf02551274"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/bf02551274", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1023250347", 
              "https://doi.org/10.1007/bf02551274"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/0005-1098(90)90044-i", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1029916917"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/0005-1098(90)90044-i", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1029916917"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/0098-1354(90)87028-n", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1038262573"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1073/pnas.79.8.2554", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1038762424"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1162/neco.1989.1.2.270", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1047038492"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1049/ip-d.1980.0047", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1056847662"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1103/physrevlett.59.2229", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1060795871"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1103/physrevlett.59.2229", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1060795871"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/21.101159", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061121293"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/37.55121", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061163219"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/72.80202", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061219259"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/72.80209", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061219266"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "1995-12-26T00:00", 
        "description": "

    The present invention is a fully connected feed forward network that includes at least one hidden layer 16. The hidden layer 16 includes nodes 20 in which the output of the node is fed back to that node as an input with a unit delay produced by a delay device 24 occurring in the feedback path 22 (local feedback). Each node within each layer also receives a delayed output (crosstalk) produced by a delay unit 36 from all the other nodes within the same layer 16. The node performs a transfer function operation based on the inputs from the previous layer and the delayed outputs. The network can be implemented as analog or digital or within a general purpose processor. Two teaching methods can be used: (1) back propagation of weight calculation that includes the local feedback and the crosstalk or (2) more preferably a feed forward gradient decent which immediately follows the output computations and which also includes the local feedback and the crosstalk. Subsequent to the gradient propagation, the weights can be normalized, thereby preventing convergence to a local optimum. Education of the network can be incremental both on and off-line. An educated network is suitable for modeling and controlling dynamic nonlinear systems and time series systems and predicting the outputs as well as hidden states and parameters. The educated network can also be further educated during on-line processing.

    ", "id": "sg:patent.US-5479571-A", "keywords": [ "network", "method", "teaching", "invention", "feed", "layer", "output", "node", "input", "delay", "feedback", "crosstalk", "transfer function", "analog", "general-purpose processor", "propagation", "calculation", "gradient", "computation", "weight", "convergence", "local optimum", "Education", "modeling", "nonlinear system", "time series", "state", "parameter", "processing" ], "name": "Neural node network and model, and method of teaching same", "recipient": [ { "id": "https://www.grid.ac/institutes/grid.264763.2", "type": "Organization" } ], "sameAs": [ "https://app.dimensions.ai/details/patent/US-5479571-A" ], "sdDataset": "patents", "sdDatePublished": "2019-04-18T10:26", "sdLicense": "https://scigraph.springernature.com/explorer/license/", "sdPublisher": { "name": "Springer Nature - SN SciGraph project", "type": "Organization" }, "sdSource": "s3://com-uberresearch-data-patents-target-20190320-rc/data/sn-export/402f166718b70575fb5d4ffe01f064d1/0000100128-0000352499/json_export_03016.jsonl", "type": "Patent" } ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/patent.US-5479571-A'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/patent.US-5479571-A'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/patent.US-5479571-A'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/patent.US-5479571-A'


     

    This table displays all metadata directly associated to this object as RDF triples.

    110 TRIPLES      15 PREDICATES      57 URIs      37 LITERALS      2 BLANK NODES

    Subject Predicate Object
    1 sg:patent.US-5479571-A schema:about anzsrc-for:2746
    2 schema:author N055f3602372d4e8c89256c1ac63e28fb
    3 schema:citation sg:pub.10.1007/bf02551274
    4 https://doi.org/10.1016/0005-1098(90)90044-i
    5 https://doi.org/10.1016/0005-1098(90)90154-a
    6 https://doi.org/10.1016/0098-1354(90)87028-n
    7 https://doi.org/10.1049/ip-d.1980.0047
    8 https://doi.org/10.1073/pnas.79.8.2554
    9 https://doi.org/10.1103/physrevlett.59.2229
    10 https://doi.org/10.1109/21.101159
    11 https://doi.org/10.1109/37.55121
    12 https://doi.org/10.1109/72.80202
    13 https://doi.org/10.1109/72.80209
    14 https://doi.org/10.1162/neco.1989.1.2.161
    15 https://doi.org/10.1162/neco.1989.1.2.270
    16 https://doi.org/10.1162/neco.1990.2.4.490
    17 schema:datePublished 1995-12-26T00:00
    18 schema:description <p>The present invention is a fully connected feed forward network that includes at least one hidden layer 16. The hidden layer 16 includes nodes 20 in which the output of the node is fed back to that node as an input with a unit delay produced by a delay device 24 occurring in the feedback path 22 (local feedback). Each node within each layer also receives a delayed output (crosstalk) produced by a delay unit 36 from all the other nodes within the same layer 16. The node performs a transfer function operation based on the inputs from the previous layer and the delayed outputs. The network can be implemented as analog or digital or within a general purpose processor. Two teaching methods can be used: (1) back propagation of weight calculation that includes the local feedback and the crosstalk or (2) more preferably a feed forward gradient decent which immediately follows the output computations and which also includes the local feedback and the crosstalk. Subsequent to the gradient propagation, the weights can be normalized, thereby preventing convergence to a local optimum. Education of the network can be incremental both on and off-line. An educated network is suitable for modeling and controlling dynamic nonlinear systems and time series systems and predicting the outputs as well as hidden states and parameters. The educated network can also be further educated during on-line processing.</p>
    19 schema:keywords Education
    20 analog
    21 calculation
    22 computation
    23 convergence
    24 crosstalk
    25 delay
    26 feed
    27 feedback
    28 general-purpose processor
    29 gradient
    30 input
    31 invention
    32 layer
    33 local optimum
    34 method
    35 modeling
    36 network
    37 node
    38 nonlinear system
    39 output
    40 parameter
    41 processing
    42 propagation
    43 state
    44 teaching
    45 time series
    46 transfer function
    47 weight
    48 schema:name Neural node network and model, and method of teaching same
    49 schema:recipient https://www.grid.ac/institutes/grid.264763.2
    50 schema:sameAs https://app.dimensions.ai/details/patent/US-5479571-A
    51 schema:sdDatePublished 2019-04-18T10:26
    52 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    53 schema:sdPublisher N26e10f3a71844bdf900e02d15cec329b
    54 sgo:license sg:explorer/license/
    55 sgo:sdDataset patents
    56 rdf:type sgo:Patent
    57 N055f3602372d4e8c89256c1ac63e28fb rdf:first N76043fc39b794374b8a24c12d4925a62
    58 rdf:rest Na9dff543219448fbb088b4d230047703
    59 N26e10f3a71844bdf900e02d15cec329b schema:name Springer Nature - SN SciGraph project
    60 rdf:type schema:Organization
    61 N2f55f480ea55441789a7ec62c125cfdb schema:name Kil T. Chong
    62 rdf:type schema:Person
    63 N3fabaa1a50c04f6396d0d1552cc56cd5 schema:name Benito Fernandez
    64 rdf:type schema:Person
    65 N525975ef2f9743439019e060c168b17b schema:name Amir F. Atiya
    66 rdf:type schema:Person
    67 N5b74ab44ddd948ed92aa2c7198d1d2e2 schema:name Wei K. Tsai
    68 rdf:type schema:Person
    69 N6a566f7c8c3340a7b858270032c36f24 rdf:first N3fabaa1a50c04f6396d0d1552cc56cd5
    70 rdf:rest N819d85cba6804c2789d2539bb7a9f94f
    71 N76043fc39b794374b8a24c12d4925a62 schema:name Alexander G. Parlos
    72 rdf:type schema:Person
    73 N819d85cba6804c2789d2539bb7a9f94f rdf:first N5b74ab44ddd948ed92aa2c7198d1d2e2
    74 rdf:rest Nab14dbf0652f43edbca6a0fcf92019fe
    75 Na9dff543219448fbb088b4d230047703 rdf:first N525975ef2f9743439019e060c168b17b
    76 rdf:rest N6a566f7c8c3340a7b858270032c36f24
    77 Nab14dbf0652f43edbca6a0fcf92019fe rdf:first N2f55f480ea55441789a7ec62c125cfdb
    78 rdf:rest rdf:nil
    79 anzsrc-for:2746 schema:inDefinedTermSet anzsrc-for:
    80 rdf:type schema:DefinedTerm
    81 sg:pub.10.1007/bf02551274 schema:sameAs https://app.dimensions.ai/details/publication/pub.1023250347
    82 https://doi.org/10.1007/bf02551274
    83 rdf:type schema:CreativeWork
    84 https://doi.org/10.1016/0005-1098(90)90044-i schema:sameAs https://app.dimensions.ai/details/publication/pub.1029916917
    85 rdf:type schema:CreativeWork
    86 https://doi.org/10.1016/0005-1098(90)90154-a schema:sameAs https://app.dimensions.ai/details/publication/pub.1001974240
    87 rdf:type schema:CreativeWork
    88 https://doi.org/10.1016/0098-1354(90)87028-n schema:sameAs https://app.dimensions.ai/details/publication/pub.1038262573
    89 rdf:type schema:CreativeWork
    90 https://doi.org/10.1049/ip-d.1980.0047 schema:sameAs https://app.dimensions.ai/details/publication/pub.1056847662
    91 rdf:type schema:CreativeWork
    92 https://doi.org/10.1073/pnas.79.8.2554 schema:sameAs https://app.dimensions.ai/details/publication/pub.1038762424
    93 rdf:type schema:CreativeWork
    94 https://doi.org/10.1103/physrevlett.59.2229 schema:sameAs https://app.dimensions.ai/details/publication/pub.1060795871
    95 rdf:type schema:CreativeWork
    96 https://doi.org/10.1109/21.101159 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061121293
    97 rdf:type schema:CreativeWork
    98 https://doi.org/10.1109/37.55121 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061163219
    99 rdf:type schema:CreativeWork
    100 https://doi.org/10.1109/72.80202 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061219259
    101 rdf:type schema:CreativeWork
    102 https://doi.org/10.1109/72.80209 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061219266
    103 rdf:type schema:CreativeWork
    104 https://doi.org/10.1162/neco.1989.1.2.161 schema:sameAs https://app.dimensions.ai/details/publication/pub.1007499213
    105 rdf:type schema:CreativeWork
    106 https://doi.org/10.1162/neco.1989.1.2.270 schema:sameAs https://app.dimensions.ai/details/publication/pub.1047038492
    107 rdf:type schema:CreativeWork
    108 https://doi.org/10.1162/neco.1990.2.4.490 schema:sameAs https://app.dimensions.ai/details/publication/pub.1003959523
    109 rdf:type schema:CreativeWork
    110 https://www.grid.ac/institutes/grid.264763.2 schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...