Contextual Search Disambiguation

Patent metadata

Field	Value
Publication number	US-2016117360-A1
Application number	US-201414524830-A
Country	US
Kind code	A1
Filing date	Oct 27, 2014
Priority date	Oct 27, 2014
Publication date	Apr 28, 2016
Grant date	—

Abstract

Methods, systems, computer-readable media, and apparatuses for providing search disambiguation using contextual information and domain ontologies are presented. In some embodiments, a computing device may receive a natural language input from a user. The computing device may identify a plurality of hypotheses for the natural language input. The computing device may map the plurality of hypotheses to one or more concepts of a plurality of concepts of an ontology by annotating the one or more concepts. The ontology may include the plurality of concepts respectively connected by a plurality of relations. The computing device may determine that there is an imperfect match between the annotated one or more concepts and annotations of answers. In response, the computing device may disambiguate the annotated one or more concepts using the ontology. The computing device may present output to the user based on the disambiguation.

First claim

We claim:

1 . A method comprising:
receiving, by a computing device, a natural language input from a user; identifying a plurality of hypotheses from the natural language input; mapping the plurality of hypotheses to one or more concepts of a plurality of concepts of an ontology by annotating the one or more concepts, wherein the ontology comprises the plurality of concepts respectively connected by a plurality of relations; in response to determining that a similarity metric measuring the similarity between the annotated one or more concepts and annotations of answers is below a predetermined threshold, disambiguating the annotated one or more concepts using the ontology; and presenting output to the user based on the disambiguation.

2 . The method of claim 1 , wherein the natural language input from the user is a verbal utterance spoken by the user.

3 . The method of claim 1 , wherein the identifying the plurality of hypotheses from the natural language input further comprises:
sending, from the computing device, the natural language input to a remote recognizer; receiving, at the computing device from the remote recognizer, a first subset of the plurality of hypotheses; and generating, by the computing device, a second subset of the plurality of hypotheses from the natural language input.

4 . The method of claim 1 , the ontology is a domain-specific ontology specific to a user manual of an appliance.

5 . The method of claim 4 , further comprising:
disambiguating the annotated one or more concepts based on a generic ontology comprising a plurality of generic concepts connected by a plurality of generic relations, wherein generic ontology is associated with an appliance type of the appliance.

6 . The method of claim 4 , further comprising:
determining, using the domain-specific ontology, that one of the annotated one or more concepts cannot be performed by the appliance or are not a part of the appliance; and disambiguating the annotated one or more concepts based on the determining.

7 . The method of claim 6 , further comprising:
determining, using a generic ontology, an alternate concept to the one of the annotated one or more concepts, wherein the output is an indication that the one of the annotated one or more concepts cannot be performed and an indication of the alternate concept.

8 . The method of claim 1 , further comprising:
in response to determining that one of the annotated one or more concepts is a superclass concept associated with a plurality of subclass concepts, determining, using the ontology, the plurality of subclass concepts; determining a current context associated with each of the plurality of subclass concepts; and disambiguating the one of the annotated one or more concepts based on the plurality of subclass concepts and the associated current context of each of the plurality of subclass concepts.

9 . The method of claim 8 , further comprising:
determining an environmental context outside of an appliance and associated with one of the plurality of subclass concepts, wherein the disambiguating the annotated one or more concepts is based on the plurality of subclass concepts, the associated current context of each of the plurality of subclass concepts, and the environmental context.

10 . The method of claim 1 , wherein the output is a list of user-selectable options.

11 . The method of claim 1 , wherein the output is an answer associated with an answer template comprising one or more answer concepts to the natural language input.

12 . The method of claim 1 , wherein the ontology comprises categories of concepts including items concepts, actions concepts, locations concepts, querytype concepts, observation concepts, answer concepts, and userquery concepts.

13 . The method of claim 1 , wherein the plurality of relations comprises at least a hasAction relation, a hasColor relation, a hasPart relation, an isAbout relation, and a hasAnswer relation.

14 . An apparatus comprising:
a processor; and a memory for storing computer readable instructions that, when executed by the processor, cause the apparatus to:
receive a natural language input from a user;
identify a plurality of hypotheses from the natural language input;
annotating one or more concepts, of a plurality of concepts, associated with one or more of the plurality of hypotheses based on an ontology comprising the plurality of concepts respectively connected by a plurality of relations;
in response to a determination that a similarity metric measuring the similarity between the annotated one or more concepts and annotations of answers does not exceed a predetermined threshold, disambiguate the annotated one or more concepts using the ontology; and
presenting output to the user based on the disambiguation.

15 . The apparatus of claim 14 , wherein the ontology is one of a domain-specific ontology for a vehicle model or a generic ontology for multiple different vehicle models.

16 . The apparatus of claim 14 , wherein the computer readable instructions that, when executed by the processor, further cause the apparatus to:
disambiguate the annotated one or more concepts using a sensor associated with one of the annotated one or more concepts.

17 . The apparatus of claim 14 , wherein the computer readable instructions that, when executed by the processor, further cause the apparatus to:
in response to a determination that one of the annotated one or more concepts is a superclass concept associated with a plurality of subclass concepts, determining, using the ontology, the plurality of subclass concepts; and disambiguating the one of the annotated one or more concepts based on the plurality of subclass concepts.

18 . A method comprising:
disambiguating, by a computing device, one or more hypotheses associated with a natural language input and mapped to one or more concepts, of a plurality of concepts, of a first ontology based on a second ontology different from the first ontology and a current context associated with one of the one or more concepts.

19 . The method of claim 18 , wherein the first ontology comprises categories of concepts including items concepts, actions concepts, locations concepts, querytype concepts, observation concepts, answer concepts, and userquery concepts.

20 . The method of claim 18 , wherein the first ontology comprises a plurality of relations including at least a hasAction relation, a hasColor relation, a hasPart relation, an isAbout relation, and a hasAnswer relation.

Assignees

• Nuance Communications Inc

Inventors

• Kunc Ladislav
• Labský Martin
• Macek Tomás
• VystrčIl Jan
• Kleindienst Jan

Classifications

• G06F16/3329 Primary

  Natural language query formulation · CPC title

• G06F40/169

  Annotation, e.g. comment data or footnotes · CPC title

• G06F16/24573

  using data annotations, e.g. user-defined metadata · CPC title

• G06F16/84

  Mapping; Conversion · CPC title

• G06F16/243 Primary

  Natural language query formulation · CPC title