http://rdf.ncbi.nlm.nih.gov/pubchem/patent/EP-3642731-A1

Outgoing Links

Predicate Object
assignee http://rdf.ncbi.nlm.nih.gov/pubchem/patentassignee/MD5_7542cdde926ccc0ad7b27ea92c083088
classificationCPCInventive http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G10L15-063
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F40-35
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06N20-00
http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G10L15-1815
classificationIPCInventive http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06K9-00
http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G10L15-00
filingDate 2018-05-30-04:00^^<http://www.w3.org/2001/XMLSchema#date>
inventor http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_eb9bf9b6003c5115fba70a685aa3ca04
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_d0c5221ec04af6e42e199ffe4338edd4
http://rdf.ncbi.nlm.nih.gov/pubchem/patentinventor/MD5_1ad7274aa5e5035dc6e9cce7aecb1af4
publicationDate 2020-04-29-04:00^^<http://www.w3.org/2001/XMLSchema#date>
publicationNumber EP-3642731-A1
titleOfInvention Intent and slot detection for digital assistants
abstract A mechanism for adapting an automatic learning model used in a language comprehension model that has been learned using a first set of user inputs including a first set of features to operate. effectively using a user input comprising a second set of features. Losses are defined according to the first set of features, the second set of features, or features common to both the first set and the second set. Losses include one or more of a source-side loss of labeling, a loss of reconstruction, a loss of an antagonistic domain classification, a loss of non-antagonistic domain classification, a loss of orthogonality and a loss of target-side marking. Losses are jointly minimized using a gradient descent method and the coefficients obtained are used to relearn the machine learning model.
isCitedBy http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-111639607-A
priorityDate 2017-07-27-04:00^^<http://www.w3.org/2001/XMLSchema#date>
type http://data.epo.org/linked-data/def/patent/Publication

Incoming Links

Predicate Subject
isDiscussedBy http://rdf.ncbi.nlm.nih.gov/pubchem/substance/SID451646703
http://rdf.ncbi.nlm.nih.gov/pubchem/compound/CID5388961

Total number of triples: 20.