Kabushiki Kaisha Toshiba

Yumiko Shimogori

Takahiro Kawamura

Tomohiro Yamasaki

Shinichi Nagano

Takayuki Iida

Hisayoshi Nagae

According to one embodiment, a classification apparatus includes the following elements. The target log extraction unit extracts a set of dialogue logs for a dialogue from dialogue logs for dialogues between a user and an interactive system, the dialogue including a first speech and a second speech of the user, the set of dialogue logs including information indicative of a classification used for a search performed by the interactive system based on the first speech, information indicative of a failure in the first search, and data acquired as a result of a search performed by the interactive system based on the second speech. The classification relation generator generates a classification relation in which the classification is associated with the data.

Classification apparatus and classification method

A data creating apparatus extracts meta data about a topic from a document, the meta data including at least one linguistic expression about a behavior, a plurality of the linguistic expressions having a first modification relation. The data creating apparatus converts the linguistic expressions included in the behavioral meta data into each class, based on a behavior ontology that is expressed by a graph where the linguistic expression about a behavior is an instance and a concept of the instance is a class to create behavior map data that represents each of the classes converted and also representing a second modification relation among the classes as a link.

Data creating apparatus and data creating method

A data-generation supporting system includes an template generating unit that generates a document type template from metadata for inputting a combination, the metadata defining a class and properties indicating attributes of an instance belonging to the class; a combination-information generating unit that receives a combination of constituent elements of the properties based on the generated document type template through an input unit, and that generates combination information; and a combination expanding unit that generates contents data using a function of calculating a direct product of the properties of the generated combination information, and using a function of performing a calculation of the constituent elements of the properties and a character-string combination of the constituent elements of the properties.

Growing community of inventors

Kanagawa, Japan

Yumiko Shimogori