Hyperbolic space-based causal inference method and system for enhancing affair atlas

A reasoning method and causal technology, applied in the field of causal reasoning, can solve the problems of lack of stability and interpretability of the model, and achieve the effect of improving stability and interpretability and reducing distortion problems.

Inactive Publication Date: 2022-05-10
HARBIN INST OF TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Through this method, a seemingly high index can be achieved. However, after verification by scholars, it is found that behind the high index is that the model fits some explic

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Hyperbolic space-based causal inference method and system for enhancing affair atlas
  • Hyperbolic space-based causal inference method and system for enhancing affair atlas
  • Hyperbolic space-based causal inference method and system for enhancing affair atlas

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0035] In order to solve the problems of lack of stability and interpretability in existing causal inference related models in the prior art, the present invention provides the following solutions:

[0036] Such as figure 1 As mentioned above, the present invention provides a causal reasoning method based on hyperbolic space-based event map enhancement, including:

[0037] Obtain the event pair and the event map, obtain the evidence events of the event pair based on the event map, and form the evidence event and the event pair into a causal evidence map, where the event pair includes the cause event and the result event;

[0038] Convert the causal evidence graph into an event representation, and map the event representation from the Euclidean space to the hyperbolic space to obtain the hyperbolic event representation; perform directed causal reasoning on the hyperbolic event representation to obtain the final embedding of the event pair, and analyze the event pair The final ...

Embodiment 2

[0048] In order to better achieve the above technical purposes, such as Figure 2-3 The shown invention provides a hyperbolic space-based causal reasoning system enhanced by the map of affairs, including:

[0049] An evidence event retrieval module and a directed hyperbolic causal reasoning model; wherein, the evidence event retrieval module is used to obtain event pairs and event graphs, obtain evidence events of event pairs based on the event graph, and form evidence events and event pairs into a causal evidence map , wherein the event pair includes a cause event and a result event; the directed hyperbolic causal inference model is used to convert the causal evidence graph into an event representation, and the event representation is mapped from a Euclidean space to a hyperbolic space to obtain a hyperbolic event representation; Directed causal inference is performed on hyperbolic event representations to obtain event-pair final embeddings, and event-pair final embeddings ar...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a hyperbolic space-based causal reasoning method and system for event atlas enhancement, and the method comprises the steps: obtaining an event pair and a event atlas, obtaining an evidence event of the event pair based on the event atlas, and forming a causal evidence graph through the evidence event and the event pair, and the event pair comprises a reason event and a result event; converting the causal evidence graph into an event representation, mapping the event representation into a hyperbolic space from an Euclidean space, and obtaining a hyperbolic event representation; and performing directed causal reasoning on the hyperbolic event representation to obtain the final embedding of the event pair, and calculating the final embedding of the event pair to obtain a causal intensity score. According to the technical scheme, the stability and interpretability of causal reasoning can be improved.

Description

technical field [0001] The invention relates to the technical field of causal reasoning, in particular to a method and system for causal reasoning based on hyperbolic space-based enhanced event map. Background technique [0002] The task of causal inference aims to enable machines to understand the various causal dependencies that are ubiquitous in real life. Such relationships exist widely in natural language texts. Many downstream applications of it have great value, such as intelligent question answering, event prediction, intelligent decision-making, etc. The specific form of the causal reasoning task is given a premise, reasoning target type (cause or result) and two possible hypotheses, requiring the machine to select a better solution among the two hypotheses as the reasoning result based on the premise. The task of causal inference has always been one of the most difficult tasks in artificial intelligence. In recent years, with the rapid development of artificial ne...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06N5/04G06N3/04
CPCG06N5/041G06N5/045G06N5/046G06N3/042G06N3/045
Inventor 丁效吴婷婷杜理秦兵刘挺王若珂
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap