Unsupervised event extraction

Patent metadata

Field	Value
Publication number	US-12001896-B2
Application number	US-202318296777-A
Country	US
Kind code	B2
Filing date	Apr 6, 2023
Priority date	May 21, 2020
Publication date	Jun 4, 2024
Grant date	Jun 4, 2024

Abstract

Computer-implemented techniques for unsupervised event extraction are provided. In one instance, a computer implemented method can include parsing, by a system operatively coupled to a processor, unstructured text comprising event information to identify candidate event components. The computer implemented method can further include employing, by the system, one or more unsupervised machine learning techniques to generate structured event information defining events represented in the unstructured text based on the candidate event components.

First claim

What is claimed is:

1. A system, comprising:
a memory that stores computer executable components;
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
a parsing component that identifies event components included in an unstructured text query using one or more unsupervised machine learning processes; and
an event extraction component that generates structured event information defining one or more events represented in the unstructured text query based on the event components.

2. The system of claim 1 , wherein the computer executable components further comprise:
a query component that queries the structured event information against structured event schema for a corpus of unstructured text data and identifies one or more parts of the unstructured text data that are relevant to a query request based on a correspondence between the structured event information and reference structured event information included in the structured event schema and associated with the one or more parts.

3. The system of claim 2 , wherein the query component further generates query result data for the unstructured text query identifying the one or more parts.

4. The system of claim 2 , wherein the event extraction component generates the structured event schema for the corpus of unstructured text using the one or more unsupervised machine learning processes.

5. The system of claim 1 , wherein the structured event information defines respective events of the one or more events using a trigger term, one or more argument terms, one or more roles of the arguments and an event type.

6. The system of claim 1 , wherein the computer executable components further comprise:
an event representation component that generates one or more event representations using graph embeddings based on the event components.

7. The system of claim 6 , wherein the event components comprise one or more event trigger terms and one or more candidate event arguments respectively associated with the one or more event trigger terms, and wherein the event representations comprise an event representation for each of the one or more event trigger terms.

8. The system of claim 6 , wherein the computer executable components further comprise:
a clustering component that employs the one or more event representations to cluster the event components into one or more event types, and wherein the event extraction component generates the structured event information based on the one or more event types and the event components respectively grouped with the one or more event types.

9. The system of claim 8 , wherein the event components comprise one or more event trigger terms and one or more event arguments respectively associated with the one or more event trigger terms, and wherein the computer executable components further comprise:
a role labeling component that labels the one or more event arguments with one or more role attributes representative of one or more roles the one or more event arguments play with respect to the one or more event types, wherein the event extraction component generates the structured event information based on the one or more event types, the one or more event trigger terms respectively associated with the one or more event types, the one or more event arguments respectively associated with the one or more event trigger terms, and the one or more role attributes.

10. The system of claim 9 , wherein the role labeling component employs one or more external knowledge bases to facilitate labeling the one or more event arguments with the one or more role attributes.

11. The system of claim 1 , wherein the one or more unsupervised machine learning processes comprise abstract meaning representation parsing.

12. A method, comprising:
identifying, by a system operatively coupled to a processor, event components included in an unstructured text query using one or more unsupervised machine learning processes; and
generating, by the system, structured event information defining one or more events represented in the unstructured text query based on the event components.

13. The method of claim 12 , further comprising:
querying, by the system, the structured event information against structured event schema for a corpus of unstructured text data; and
identifying, by the system, one or more parts of the unstructured text data that are relevant to a query request based on a correspondence between the structured event information and reference structured event information included in the structured event schema and associated with the one or more parts.

14. The method of claim 13 , further comprising:
generating, by the system, query result data for the unstructured text query identifying the one or more parts.

15. The method of claim 12 , further comprising:
generating, by the system, the structured event schema for a corpus of unstructured text using the one or more unsupervised machine learning processes.

16. The method of claim 12 , wherein the structured event information defines respective events of the one or more events using a trigger term, one or more argument terms, one or more roles of the arguments and an event type.

17. The method of claim 12 , further comprising:
generating, by the system, one or more event representations using graph embeddings based on the event components;
employing, by the system, the one or more event representations to cluster the event components into one or more event types; and
generating, by the system, the structured event information based on the one or more event types and the event components respectively grouped with the one or more event types.

18. The method of claim 17 , wherein the event components comprise one or more event trigger terms and one or more event arguments respectively associated with the one or more event trigger terms, and wherein the method further comprises:
labeling, by the system, the one or more event arguments with one or more role attributes representative of one or more roles the one or more event arguments play with respect to the one or more event types; and
generating, by the system, the structured event information based on the one or more event types, the one or more event trigger terms respectively associated with the one or more event types, the one or more event arguments respectively associated with the one or more event trigger terms, and the one or more role attributes.

19. A computer program product for unsupervised event extraction, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processing component to cause the processing component to:
identify event components included in an unstructured text query using one or more unsupervised machine learning processes; and
generate structured event information defining one or more events represented in the unstructured text query based on the event components.

20. The computer program product of claim 19 , the program instructions further executable by the processing component to cause the processing component to:
query the structured event information against structured event schema for a corpus of unstructured text data; and
identify one or more parts of the unstructured text data that are relevant to a query request based on a correspondence between the structured event information and reference structured event inf

Assignees

• IBM

Inventors

• Haldar Rajarshi
• Deng Yu
• Wu Lingfei
• Mahindru Ruchi
• Tao Shu

Classifications

• G06F9/542 Primary

  Event management; Broadcasting; Multicasting; Notifications · CPC title

• G06F16/35

  Clustering; Classification · CPC title

• G06F16/38

  Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually · CPC title

• G06F16/9024

  Graphs; Linked lists (G06F16/9027 takes precedence) · CPC title

• G06F40/205

  Parsing · CPC title