Open microphone perpetual conversation analysis

Patent metadata

Field	Value
Publication number	US-9710460-B2
Application number	US-201514735285-A
Country	US
Kind code	B2
Filing date	Jun 10, 2015
Priority date	Jun 10, 2015
Publication date	Jul 18, 2017
Grant date	Jul 18, 2017

Abstract

A conversation analytics system including multiple microphones that each respectively capture at least a part of a single real world natural language conversation. The natural language data (for example, audio of natural language, or audio of natural language converted to text) from each of the multiple microphones is combined into a single combined piece of natural language data (for example, a combined audio file, or a combined piece of text). This combined piece of natural language data is subject to: (i) machine logic based natural language understanding; and/or (ii) community conversation analysis. The multiple microphones may be respectively built into mobile devices which are always on and which are generally always carried by the device owners on their respective persons.

First claim

What is claimed is:

1. A method comprising:
determining, based on a context of a conversation, that a privacy policy is not triggered, wherein:
the privacy is triggered based on receiving a signal; and
the privacy policy prevents receiving audio data;

receiving first natural language data based on audio data sensed by a first microphone and second natural language data based on audio data sensed by a second microphone, where the first microphone and the second microphone are within conversational proximity to each other when the first microphone and the second microphone respectively sense the first natural language data and the second natural language data, wherein the first microphone senses natural language of at least a first user and a second user, and wherein the second microphone senses natural language of at least the first user and the second user;
determining, by machine logic, that each of the first natural language data and the second natural language data correspond to a first real world natural language conversation between the first user and the second user;
determining, by machine logic, that the first natural language data and the second natural language data corresponding to the first real world natural language conversation include at least a first conversational portion that does not match;
reconciling, by machine logic, at least the first conversational portion to yield at least a first matching conversational portion;
combining the first natural language data and the second natural language data into a third natural language data corresponding to the natural language of the first real world natural language conversation, wherein the third natural language data includes at least the first matching conversational portion; and
performing, by machine logic, natural language understanding on the third natural language data to yield at least a first insight information.

2. The method of claim 1 further comprising:
outputting the first insight information in human understandable form and format.

3. The method of claim 1 further comprising:
performing conversation analytics on the third natural language data to determine at least one of the following: sentiment, intent, social leader identification, propagation tracking, common conversation identification, and/or visualization.

4. The method of claim 3 wherein:
the performance of conversation analytics includes propagation tracking; and subject matter of the propagation tracking includes an idea.

5. The method of claim 1 wherein the first natural language data and the second natural language data are received as text-based data derived from audio sensed by the first microphone and the second microphone.

6. The method of claim 1 wherein:
the first microphone is built into a first mobile device which is always on; and
the second microphone is built into a second mobile device which is always on.

7. A computer program product comprising a computer readable storage medium having stored thereon:
first program instructions programmed to determine, based on a context of a conversation, that a privacy policy is not triggered, wherein:
the privacy is triggered based on receiving a signal; and
the privacy policy prevents receiving audio data;

second program instructions programmed to receive first natural language data based on audio data sensed by a first microphone and second natural language data based on audio data sensed by a second microphone, where the first microphone and the second microphone are within conversational proximity to each other when the first microphone and the second microphone respectively sense the first natural language data and the second natural language data, wherein the first microphone senses natural language of at least a first user and a second user, and wherein the second microphone senses natural language of at least the first user and the second user;
third program instructions programmed to determine, by machine logic, that each of the first natural language data and the second natural language data correspond to a first real world natural language conversation between the first user and the second user;
fourth program instructions programmed to determine, by machine logic, that the first natural language data and the second natural language data corresponding to the first real world natural language conversation include at least a first conversational portion that does not match;
fifth program instructions programmed to reconcile, by machine logic, at least the first conversational portion to yield at least a first matching conversational portion;
sixth program instructions programmed to combine the first natural language data and the second natural language data into a third natural language data corresponding to the natural language of the first real world natural language conversation, wherein the third natural language data includes at least the first matching conversational portion; and
seventh program instructions programmed to perform, by machine logic, natural language understanding on the third natural language data to yield at least a first insight information.

8. The product of claim 7 wherein the medium has further stored thereon:
eighth program instructions programmed to output the first insight information in human understandable form and format.

9. The product of claim 7 wherein the medium has further stored thereon:
ninth program instructions programmed to perform conversation analytics on the third natural language data to determine at least one of the following: sentiment, intent, social leader identification, propagation tracking, common conversation identification, and/or visualization.

10. The product of claim 9 wherein:
the ninth program instructions are further programmed to perform conversation analytics including propagation tracking; and
subject matter of the propagation tracking includes an idea.

11. The product of claim 7 wherein the first program instructions are further programmed to receive the first natural language data and the second natural language data as text-based data derived from audio sensed by the first microphone and the second microphone.

12. The product of claim 7 wherein:
the first microphone is built into a first mobile device which is always on; and the second microphone is built into a second mobile device which is always on.

13. The product of claim 7 further comprising a computer system including:
a processor(s) set; and
a computer readable storage medium; wherein:
the processor set is structured, located, connected and/or programmed to run program instructions stored on the computer readable storage medium.

14. A method comprising:
determining, based on a context of a conversation, that a privacy policy is not triggered, wherein:
the privacy is triggered based on receiving a signal; and
the privacy policy prevents receiving audio data;

receiving first natural language data based on audio data sensed by a first microphone, deployed in a first community, and second natural language data based on audio data sensed by a second microphone, deployed in the first community, where the first microphone and the second microphone are within conversational proximity to each other when the first microphone and the second microphone respectively sense the first natural language data and the second natural language data, wherein the first microphone senses natural language of at least a first user and a second user, and wherein the second microphone senses natural language of at least the first user and the second user;
determining, by machine

Assignees

• IBM

Inventors

• Brown Michael S
• Harrer Stefan
• Plant Laurence J

Classifications

• G10L15/32

  Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems · CPC title

• G06F40/35 Primary

  Discourse or dialogue representation · CPC title

• G06F17/279 Primary

  Physics · mapped topic