公开(公告)号：US20210064825A1

公开(公告)日：2021-03-04

申请号：US16961079

申请日：2019-02-06

Applicant: NEC CORPORATION

Inventor： Kunihiko SADAMASA ,  Yuzuru OKAJIMA

IPC: G06F40/30 ,  G06F40/279 ,  G06N5/04 ,  G06N20/00 ,  G06F16/2455

Abstract: An information processing apparatus 10 includes an acceptance unit 12, a formal language generation unit 14, an inference unit 16, and an update unit 18. The formal language generation unit 14 generates training expressions in a formal language based on parameters prepared in advance and pieces of text accepted by the acceptance unit 12. The inference unit 16 executes at least one inference out of deductive inference and abduction on the above training expressions using a knowledge base prepared in advance. The update unit 18 compares an inference result of the inference unit 16 with a model answer input in advance, and updates parameters such that an inference result that matches the model answer is likely to be obtained through inference performed by the inference unit 16.

公开(公告)号：US20180314951A1

公开(公告)日：2018-11-01

申请号：US15772678

申请日：2015-11-10

Applicant: NEC CORPORATION

Inventor： Kunihiko SADAMASA ,  Takashi ONISHI ,  Kentarou SASAKI ,  Yotaro WATANABE ,  Kai ISHIKAWA ,  Satoshi MORINAGA

IPC: G06N5/02 ,  G06N5/04

CPC classification number: G06N5/04

Abstract: A reasoning system that enables reasoning when there is a shortage of knowledge. An input unit receives a start state and an end state. A rule candidate generation unit identifies a first state, obtained by tracking one or more known rules from the start state, and a second state, obtained by backtracking one or more known rules from the end state, respectively. The generation unit generates a rule candidate relating to the first state and the second state or generates a rule candidate relating to the first state and a rule candidate relating to the second state. A rule selection unit selects, based on feasibility of the generated rule candidate, which is calculated based on one or more known rules, the generated rule candidate as a new rule. A derivation unit derives the end state from the start state, based on one or more known rules and the new rule.

公开(公告)号：US20220245183A1

公开(公告)日：2022-08-04

申请号：US17613549

申请日：2019-05-31

Applicant: NEC Corporation

Inventor： Kosuke AKIMOTO ,  Takuya HIRAOKA ,  Kunihiko SADAMASA

IPC: G06F16/33 ,  G06N3/04

Abstract: An entity combination and a text representation are obtained as a first fact, and the entity combination and a related predicate are obtained as a second fact. Word distributed representations are input to a neural network, real vectors at appearance positions of entities are specified and used as distributed representations. A first score indicating a degree of establishment of the first fact is calculated based on the distributed representations and on entity distributed representations. A second score indicating a degree of establishment is calculated with respect to an entity combination and a text representation that are not the first fact. A third score indicating a degree of establishment of the second fact is calculated based on predicate distributed representations and on entity distributed representations. A fourth score indicating a degree of establishment is calculated also with respect to an entity combination and a predicate that are not the second fact. The entity distributed representations, the predicate distributed representations, or weight parameters are updated by a gradient method, so that the first score becomes higher than one of the second score and the fourth score, and the third score becomes higher than one of the second score and the fourth score.

公开(公告)号：US20210209447A1

公开(公告)日：2021-07-08

申请号：US17058809

申请日：2018-05-31

Applicant: NEC Corporation

Inventor： Yuzuru OKAJIMA ,  Kunihiko SADAMASA

IPC: G06N3/04 ,  G06N5/00 ,  G06N5/02

Abstract: An information processing apparatus (2000) acquires input data (10). The information processing apparatus (2000) extracts a prediction rule (50) used for prediction related to the input data (10) from a usage rule set (60) by using a neural network (30). The usage rule set (60) includes a plurality of candidates for the prediction rule (50) used for prediction related to the input data (10). The prediction rule (50) is information in which condition data (52) representing a basis for prediction and conclusion data (54) representing a prediction related to the input data (10) are associated with each other. The prediction rule (50) used for prediction related to the input data (10) indicates the condition data (52) indicating a condition satisfied by the input data (10). The information processing apparatus (2000) outputs a prediction result (20), based on the conclusion data (54) indicating the extracted prediction rule (50).

公开(公告)号：US20200293929A1

公开(公告)日：2020-09-17

申请号：US16083992

申请日：2017-03-09

Applicant: NEC Corporation

Inventor： Kentarou SASASKI ,  Daniel Georg ANDRADE SILVA ,  Yotaro WATANABE ,  Kunihiko SADAMASA

IPC: G06N5/04 ,  G06N20/00 ,  G06K9/62 ,  G06F17/18

Abstract: An inference method according to the present invention in an inference system inferring a probability that an ending state holds based on a starting state and a rule set, the method includes: when a rule set derived by excluding one rule from rules constituting a first rule set is set as a second rule set, a probability that the ending state holds based on the starting state and the first rule set is set as a first inference result, and a probability that the ending state holds based on the starting state and the second rule set is set as a second inference result, calculating an importance being an indicator indicating magnitude of a difference between the first inference result and the second inference result; and outputting the rule and the importance of the rule, being associated with each other for each of the excluded rule.

公开(公告)号：US20200057762A1

公开(公告)日：2020-02-20

申请号：US16487912

申请日：2017-03-31

Applicant: NEC Corporation

Inventor： Yuzuru OKAJIMA ,  Kunihiko SADAMASA

IPC: G06F16/242 ,  G06F16/2452 ,  G06F16/2453 ,  G06N20/00

Abstract: This invention provides an information processing apparatus for learning parameters to convert a question in a natural language into a query expression in a formal language, including an acceptor accepting an input of a set of a question in the natural language and a correct answer to the question, a condition generator generating condition(s) to be satisfied by the formal language input on searching a database for the correct answer, a query expression generator generating a query expression in the formal language corresponding to the question using parameters of the converter to satisfy any of the at least one condition, an answer acquirer acquiring an answer to the question based on a search of the database using the generated query expression, and a updater updating the parameters such that the question is converted with priority into the generated query expression when the answer and the correct answer match.

公开(公告)号：US20190180192A1

公开(公告)日：2019-06-13

申请号：US16323285

申请日：2016-08-18

Applicant: NEC Corporation

Inventor： Daniel Georg ANDRADE SILVA ,  Yotaro WATANABE ,  Satoshi MORINAGA ,  Kunihiko SADAMASA

IPC: G06N5/02 ,  G06F16/901 ,  G06N7/00

Abstract: An information processing system for learning new probabilistic rules even if only one training sample is given. A learning system (100) includes a KB (knowledge base) storage (110), a rule generator (130), and a weight calculator (140). The KB storage (110) stores a KB including a knowledge storage for storing rules between events among a plurality of events. The rule generator (130) generates one or more new rules based on the rules and an implication score between the events. The weight calculator (140) calculates a weight of the one or more new rules for probabilistic reasoning based on the implication score.