Automated timeline completion using event progression knowledge base

Patent metadata

Field	Value
Publication number	US-10372822-B2
Application number	US-201615172813-A
Country	US
Kind code	B2
Filing date	Jun 3, 2016
Priority date	Jun 3, 2016
Publication date	Aug 6, 2019
Grant date	Aug 6, 2019

Abstract

A mechanism is provided in a computing device configured with instructions executing on a processor of the computing device to implement a timeline generation system, for automated timeline completion. The timeline generation system executing on the processor of the computing device identifies a plurality of events in documents in a corpus of information. The timeline generation system places the plurality of events in a partial timeline data structure. The timeline generation system selects an event progression from an event progression knowledge base. The timeline generation system aligns the selected event progression to the partial timeline data structure. The timeline generation system identifies a set of events missing from the partial timeline data structure. The timeline generation system maps the set of events missing from the partial timeline data structure to the partial timeline based on the selected event progression to form a completed timeline data structure.

First claim

What is claimed is:

1. A method, in a computing device configured with instructions executing on a processor of the computing device to implement a natural language processing system, for automated timeline completion, the method comprising:
identifying, by a timeline generation component executing within a natural language processing pipeline of the natural language processing system, a plurality of events having time information in documents in a corpus of information;
placing, by the timeline generation component, the plurality of events in a partial timeline data structure;
selecting, by the timeline generation component, an event progression from an event progression knowledge base, wherein the event progression knowledge base is a knowledge resource consisting of multiple event progressions, wherein each event progression in the event progression knowledge base is a natural ordering of a sequence of events as they typically occur in the real world, and wherein each event in a given event progression has the following time information: rank information implicit in the ordering, approximate distance from the next event, and confidence or fuzziness factor for the distance, wherein selecting the event progression from the event progression knowledge base comprises:
aligning each event progression in the event progression knowledge base to the partial timeline data structure;
determining a matching score of each event progression based on a mapping function; and
selecting an event progression having a highest matching score;

aligning, by the timeline generation component, the selected event progression to the partial timeline data structure;
identifying, by the timeline generation component, a set of events present in the selected event progression and missing from the partial timeline data structure;
identifying, by the timeline generation component, a set of unassigned events in the documents in the corpus that lack associated time information and are not placed the partial timeline data structure;
matching, by the timeline generation component, the set of unassigned events to the selected event progression based on semantic similarity or synonymity between event pairs to form a set of matched unassigned events;
mapping, by the timeline generation component, the set of matched events to the partial timeline based on the selected event progression to form a completed timeline data structure including the plurality of events having time information and the set of matched events;
storing, by the timeline generation component, the completed timeline data structure in the corpus of information; and
performing by the natural language processing system, a time-based natural language processing operation using the completed timeline data structure and the corpus of information.

2. The method of claim 1 , wherein identifying the plurality of events comprises performing natural language processing on unstructured documents in the corpus of information.

3. The method of claim 1 , wherein the mapping function outputs a similarity score between events across the event progression and the partial timeline data structure.

4. The method of claim 3 , wherein the mapping function outputs a similarity score of 1 if an event in the event progression and an event in the partial timeline data structure are identical or synonymous and outputs a similarity score of 0 if the event in the event progression and the event in the partial timeline data structure are not identical or synonymous.

5. The method of claim 4 , wherein the mapping function determines a similarity score using a knowledge-based approach based on an ontology data structure.

6. The method of claim 1 , wherein aligning each event progression in the event progression knowledge base to the partial timeline data structure comprises aligning each event progression in the event progression knowledge base to the partial timeline data structure using a dynamic programming algorithm.

7. The method of claim 1 , wherein mapping the set of events missing from the partial timeline data structure to the partial timeline comprises matching the set of events missing from the partial timeline data structure to the selected event progression based on semantic similarity or synonymity between event pairs.

8. The method of claim 7 , wherein semantic similarity is calculated using a knowledge-based approach based on an ontology data structure.

9. The method of claim 1 , wherein performing the time-based processing comprises performing question answering in a natural language system pipeline executing within the natural language processing system.

10. A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program comprises instructions, which when executed on a processor of a computing device causes the computing device to implement a timeline generation system for automated timeline completion, wherein the computer readable program causes the computing device to:
identify, by a timeline generation component executing within a natural language processing pipeline of the natural language processing system, a plurality of events having time information in documents in a corpus of information;
place, by the timeline generation component, the plurality of events in a partial timeline data structure;
select, by the timeline generation component, an event progression from an event progression knowledge base, wherein the event progression knowledge base is a knowledge resource consisting of multiple event progressions, wherein each event progression in the event progression knowledge base is a natural ordering of a sequence of events as they typically occur in the real world, and wherein each event in a given event progression has the following time information: rank information implicit in the ordering, approximate distance from the next event, and confidence or fuzziness factor for the distance, wherein selecting the event progression from the event progression knowledge base comprises:
aligning each event progression in the event progression knowledge base to the partial timeline data structure;
determining a matching score of each event progression based on a mapping function; and
selecting an event progression having a highest matching score;

align, by the timeline generation component, the selected event progression to the partial timeline data structure;
identify, by the timeline generation component, a set of events present in the selected event progression and missing from the partial timeline data structure;
identify, by the timeline generation component, a set of unassigned events in the documents in the corpus that lack associated time information and are not placed the partial timeline data structure;
match, by the timeline generation component, the set of unassigned events to the selected event progression based on semantic similarity or synonymity between event pairs to form a set of matched unassigned events;
map, by the timeline generation component, the set of matched events to the partial timeline based on the selected event progression to form a completed timeline data structure including the plurality of events having time information and the set of matched events;
store, by the timeline generation component, the completed timeline data structure in the corpus of information; and
perform, by the natural language processing system, a time-based natural language processing operation using the completed timeline data structure and the corpus of information.

11. The computer program product of claim 10 , wherein the mapping function outputs a

Assignees

• IBM

Inventors

• Devarakonda Murthy V
• Patwardhan Siddharth A
• Raghavan Preethi

Classifications

• G06F16/5866

  using information manually generated, e.g. tags, keywords, comments, manually generated location and time information · CPC title

• G06F16/447 Primary

  Temporal browsing, e.g. timeline · CPC title

• G16H10/00

  ICT specially adapted for the handling or processing of patient-related medical or healthcare data (for medical reports G16H15/00; for therapies or health-improving plans G16H20/00; for the handling or processing of medical images G16H30/00) · CPC title

• G06F40/30 Primary

  Semantic analysis · CPC title

• G16H10/60

  for patient-specific data, e.g. for electronic patient records · CPC title