Graph Neural Network for Causal Relationship Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing causality determination methods rely on artificially constructed semantic features, leading to inaccurate results, high labor costs, and low efficiency in determining causal relationships between events.
Innovation Solution
A method utilizing a graph neural network to obtain event words and related words from a target text, converting them into semantic vectors, and determining causal relationships through a convolutional and fully connected layer, thereby accurately identifying causal connections while reducing labor costs and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If artificially constructed semantic features are used to determine causal relationships between events, then the determination process can be performed, but the accuracy of causal relationship determination deteriorates
Solution Approach 1:
The patent replaces the manual mechanical process of constructing semantic features with an automated neural network system. The graph neural network automatically learns and extracts semantic features from event descriptions, eliminating the need for人工 construction and significantly improving both accuracy and reliability in causal relationship determination.
Solution Approach 2:
The patent changes the parameter representation from manually constructed semantic features to automatically learned vector representations through neural networks. This parameter transformation enables the system to capture more nuanced semantic information and improve the precision of causal relationship detection.
2Productivity
If artificially constructed semantic features are used for causality determination, then the process can be completed, but labor costs increase
Solution Approach 1:
The patent implements self-service by enabling the system to automatically construct and process semantic features without human intervention. The graph neural network autonomously extracts features from event descriptions and performs causal relationship determination, eliminating the need for manual feature engineering and significantly reducing labor costs while improving efficiency.
Solution Approach 2:
The patent substitutes the manual mechanical process of semantic feature construction with an automated neural network system, replacing human labor with intelligent algorithms that can process events at scale without additional labor costs.
3Measurement precision
If artificially constructed semantic features are used to determine causal relationships, then the method can be applied, but determination efficiency deteriorates
Solution Approach 1:
The patent replaces the inefficient manual process of constructing and analyzing semantic features with a automated graph neural network system. This substitution maintains high accuracy in causal relationship detection while dramatically improving determination efficiency through automated feature extraction and processing.
Solution Approach 2:
The patent performs preliminary action by pre-training the graph neural network on large corpora to learn effective semantic representations before actual causal relationship determination. This preliminary learning phase enables the system to quickly and accurately determine causal relationships without manual feature construction during the actual determination process.
Data Source
AI summary
Embodiments of the present disclosure provide a method for determining causality, an apparatus for determining causality, an electronic device and a storage medium, and relates to a field of knowledge graph technologies. The method includes: obtaining event words expressing individual events and related words adjacent to the event words in a target text; inputting the event words and the related words into a graph neural network; and determining whether there is a causal relationship between any two events through the graph neural network.