Identifying spans using visual recognition

Patent metadata

Field	Value
Publication number	US-11120215-B2
Application number	US-201916393145-A
Country	US
Kind code	B2
Filing date	Apr 24, 2019
Priority date	Apr 24, 2019
Publication date	Sep 14, 2021
Grant date	Sep 14, 2021

Abstract

Aspects of the present disclosure relate to identifying spans within unstructured electronic text. Natural language content is received. A part of speech and slot name of each word within the natural language content is identified. A parse tree representation is then generated based on the natural language content, wherein visual characteristics of each node of a plurality of nodes within the parse tree representation depend on the part of speech and slot name of each word. A bounding box identifying a span category is then generated around a set of nodes on the parse tree representation by a machine learning model.

First claim

What is claimed is:

1. A computer-implemented method comprising:
receiving natural language content;
identifying, by analyzing the natural language content using a natural language processor, a part of speech and slot name of respective words within the natural language content;
generating a parse tree representation based on the natural language content, wherein visual characteristics of respective nodes of a plurality of nodes within the parse tree representation depend on the part of speech and slot name of the respective words;
generating, by a machine learning model, a bounding box identifying a span category around a set of nodes on the parse tree representation, wherein the machine learning model is trained using supervised learning input data of pre-classified parse tree representations;
mapping the set of nodes encompassed by the bounding box to a set of words in the natural language content; and
outputting the set of words and the span category to a graphical user interface (GUI).

2. The method of claim 1 , wherein the machine learning model utilizes a region-based convolutional neural network.

3. The method of claim 1 , wherein each unique part of speech is associated with a unique shape, and wherein each unique slot name is associated with a unique texture.

4. The method of claim 1 , wherein the span category is factual or hypothetical.

5. The method of claim 1 , wherein the span category is selected from a group consisting of: planned event, negation, family history, and factual event.

6. A system comprising:
a computer-readable storage medium storing instructions, which, when executed by a processor, is configured to cause the processor to perform a method comprising:
receiving natural language content;
identifying, by analyzing the natural language content using a natural language processor, a part of speech and slot name of respective words within the natural language content;
generating a parse tree representation based on the natural language content, wherein visual characteristics of respective nodes of a plurality of nodes within the parse tree representation depend on the part of speech and slot name of the respective words;
generating, by a machine learning model, a bounding box identifying a span category around a set of nodes on the parse tree representation, wherein the machine learning model is trained using supervised learning input data of pre-classified parse tree representations;
mapping the set of nodes encompassed by the bounding box to a set of words in the natural language content; and
outputting the set of words and the span category to a graphical user interface (GUI).

7. The system of claim 6 , wherein the machine learning model utilizes a region-based convolutional neural network.

8. The system of claim 6 , wherein each unique part of speech is associated with a unique shape, and wherein each unique slot name is associated with a unique texture.

9. The system of claim 6 , wherein the span category is factual or hypothetical.

10. The system of claim 6 , wherein the span category is selected from a group consisting of: planned event, negation, family history, and factual event.

11. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions executable by a processor to cause the processor to perform a method comprising:
receiving natural language content;
identifying, by analyzing the natural language content using a natural language processor, a part of speech and slot name of respective words within the natural language content;
generating a parse tree representation based on the natural language content, wherein visual characteristics of respective nodes of a plurality of nodes within the parse tree representation depend on the part of speech and slot name of the respective words;
generating, by a machine learning model, a bounding box identifying a span category around a set of nodes on the parse tree representation, wherein the machine learning model is trained using supervised learning input data of pre-classified parse tree representations;
mapping the set of nodes encompassed by the bounding box to a set of words in the natural language content; and
outputting the set of words and the span category to a graphical user interface (GUI).

12. The computer program product of claim 11 , wherein the machine learning model utilizes a region-based convolutional neural network.

13. The computer program product of claim 11 , wherein each unique part of speech is associated with a unique shape, and wherein each unique slot name is associated with a unique texture.

14. The computer program product of claim 11 , wherein the span category is factual or hypothetical.

Assignees

• IBM

Inventors

• Mwarabu Chris
• Contreras David
• Delima Roberto
• Allen Corville O

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06F40/211

  Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars · CPC title

• G06F40/284

  Lexical analysis, e.g. tokenisation or collocates · CPC title