Automated Knowledge Graph Relationship Selection via Prediction Error
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual selection of relationships for knowledge graphs is time-consuming, error-prone, and biased, making it challenging to create effective cognitive computing applications that require concise, higher-level relationships for decision-making and question-answering tasks.
Innovation Solution
A computer-implemented method that identifies relationships in text documents, builds predictive models, and determines whether to store these relationships in memory based on prediction error, using techniques such as directed acyclic graphs and conciseness measures to automatically select and augment knowledge graphs with higher-level relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of relationships is used for knowledge graphs, then relationships can be carefully chosen, but the process is time-consuming and error-prone
Solution Approach 1:
The system performs automatic relationship selection using predictive models and evaluation metrics, eliminating the need for manual human intervention in the relationship selection process while maintaining or improving selection quality through algorithmic evaluation
Solution Approach 2:
The patent replaces the manual mechanical process of relationship selection with an automated computational system that uses predictive models, error calculation, and threshold-based decision making to select relationships for the knowledge graph
2Adaptability or versatility
If manual relationship selection is used, then bias can be incorporated, but the process becomes error-prone
Solution Approach 1:
The system calculates prediction error as feedback to evaluate the quality of relationships, using this error metric to objectively determine whether relationships meet the threshold for inclusion, thereby reducing subjective bias while maintaining reliability through measurable criteria
Solution Approach 2:
The patent changes the selection criterion from subjective human judgment to an objective parameter-based approach using prediction error thresholds, transforming the relationship selection process into a parameter-driven automated decision system
3Quantity of substance
If comprehensive relationships are included in knowledge graphs, then coverage is improved, but conciseness is reduced
Solution Approach 1:
The system uses prediction error as a parameter to filter and select relationships, changing the selection criterion from quantity-based to quality-based, thereby including only relationships that meet the error threshold and maintaining conciseness while achieving sufficient coverage
Data Source
AI summary
A computer-implemented method of analyzing text documents, includes identifying a relationship in a text document associated with an entity, building a predictive model from training data, in response to said identifying a relationship, wherein the predictive model includes a prediction error, and determining whether to store the identified relationship in memory, based on the prediction error.