Mining procedure dialogs from source content

Patent metadata

Field	Value
Publication number	US-10037362-B1
Application number	US-201715841421-A
Country	US
Kind code	B1
Filing date	Dec 14, 2017
Priority date	Jul 24, 2017
Publication date	Jul 31, 2018
Grant date	Jul 31, 2018

Abstract

An embodiment of the invention may include a method, computer program product and computer system for human-machine communication. The method, computer program product and computer system may include a computing device that maps linguistic data of source content to a vector. The computing device may cluster the linguistic data of source content. The computing device may determine a plurality of segments based on the mapped linguistic data and the clustered linguistic data. The computing device may transform a segment of the plurality of segments into representative data, the representative data is a function of the remaining plurality of segments.

First claim

What is claimed is:

1. A method for human-machine communication, the method comprising:
receiving source content, the source content comprising linguistic data, from one or more databases at a server communicating with the one or more databases using a communication network;
embedding the linguistic data of the source content to a high-dimensional vector using a neural network;
clustering the sentences of the source content into a plurality of sentence groups based on the embedded linguistic data;
sequencing the clustered sentence groups of the source content into a set of sentence sequences, wherein each of the plurality of sentence groups is represented by a single representative sentence;
transforming each sentence of the source content by replacing each sentence with the single representative sentence of the sentence group that each sentence is sequenced into;
embedding each representative sentence of the source content based on the sentence that precedes and follows each sentence;
mapping the relationships between the embedded representative sentences of the source content;
identifying related representative sentences of the source content;
generating linguistic data blocks comprised of the related representative sentences of the source content;
identifying similarities between the linguistic data blocks based on cohesions;
separating the linguistic data blocks into groups based on the cohesions identified; and
generating a procedure dialogue based on the cohesions between data blocks of the same group.

Assignees

• IBM
• Int Business Machines Corpoation

Inventors

• Li qi cheng
• Mei Li Jun
• Wang Jian
• Yu yi peng
• Zhou Xin

Classifications

• G06F16/90332

  Natural language query formulation or dialogue systems · CPC title

• G06F40/289 Primary

  Phrasal analysis, e.g. finite state techniques or chunking · CPC title

• G06F18/2411

  based on the proximity to a decision surface, e.g. support vector machines · CPC title

• G06F16/2428

  Query predicate definition using graphical user interfaces, including menus and forms (G06F16/2423 takes precedence) · CPC title

• G06F16/172

  Caching, prefetching or hoarding of files · CPC title