-
公开(公告)号:US10817551B2
公开(公告)日:2020-10-27
申请号:US15943742
申请日:2018-04-03
Inventor: Koji Morikawa , Yuki Minoda , Asuka Sakai
IPC: G06F16/00 , G06F16/332 , G06F16/903 , G06F40/30
Abstract: A method for expanding a word performed by a processor includes (a) obtaining a first word, (b) obtaining, from a memory, a concept map that is unique to a user and that includes a plurality of second words and semantic distances between the plurality of second words, the plurality of second words including words belonging to different categories, (c) determining, if the plurality of second words include a word corresponding to the first word, the word corresponding to the first word as a basic word, (d) selecting, on the basis of the semantic distances to the basic word, at least one of the plurality of second words included in the concept map as an adjacent word, and (e) outputting a result of a search on the basis of the first word and the adjacent word.
-
公开(公告)号:US10860800B2
公开(公告)日:2020-12-08
申请号:US16133803
申请日:2018-09-18
Inventor: Asuka Sakai , Mitsuru Endo , Takashi Ushio , Hongjie Shi
IPC: G06F40/295 , G06N20/00
Abstract: An information processing method includes acquiring first text information from a storage apparatus in which the first text information representing one or more utterance sentences is stored as a learning data set, identifying one or more named entities included in the acquired first text information, replacing each of the one or more identified named entities with an abstract expression abstracted based on a predetermined rule thereby generating second text information from the first text information, and learning a model of a dialogue system using, as training data, the second text information generated in the replacing.
-
3.
公开(公告)号:US10984026B2
公开(公告)日:2021-04-20
申请号:US15943743
申请日:2018-04-03
Inventor: Yuki Minoda , Koji Morikawa , Asuka Sakai
IPC: G06F16/332 , G06F3/01 , G06F40/30 , G06F40/284
Abstract: A method includes (a) obtaining a search word, (b) obtaining first to third concept maps including words and semantic distances between the words, (c) obtaining a first association map including degrees of association indicating how close the semantic distances included in the first and second concept maps are to each other; (d) obtaining a second association map including degrees of association indicating how close the semantic distances included in the first to third concept maps are to one another, (e) extracting, from the words as an associated word, at least one word whose difference between the degree of association with the search word included in the first association map and the degree of association with the search word included in the second association map is equal to or larger than a first threshold, and (f) outputting a result of a search based on the search word and the associated word.
-
-