Query Triplet Classification via Reinforcement Learning Arguments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Knowledge graphs used in search engines and fact-checking systems often contain incorrect and incomplete data, making it difficult to accurately classify query triplets as true or false, especially in mission-critical systems where explainability and transparency are essential.
Innovation Solution
A computer-implemented method using reinforcement learning to extract affirmative and opposing arguments from knowledge graphs, followed by supervised machine learning to classify query triplets as true or false, providing interpretable and reliable results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hand crafted rules are applied on knowledge graphs, then fact checking can be performed, but the rules become very complex and inapplicable on large amounts of data
Solution Approach 1:
The patent replaces hand-crafted mechanical rule-based systems with machine learning models that automatically learn patterns from data. The system uses trained classifiers and neural networks to perform fact-checking without requiring explicit human-defined rules, thereby reducing rule complexity while maintaining reliability.
Solution Approach 2:
The system enables the knowledge graph to self-evaluate facts through automated machine learning models. The models independently analyze triples and determine their truthfulness without requiring external manual rule application, allowing the system to serve itself in performing fact-checking at scale.
2Reliability
If hand crafted rules are applied on knowledge graphs, then fact checking can be performed, but the rules do not provide human interpretable justifications
Solution Approach 1:
The patent introduces attention mechanisms and explanation modules as intermediaries between the machine learning models and the final classification. These components generate human-interpretable justifications by highlighting which parts of the knowledge graph influenced the decision, bridging the gap between automated processing and explainability.
3Reliability
If manual inspection of data by experts is performed, then erroneous data can be identified and corrected, but the task becomes difficult and time-consuming
Solution Approach 1:
The patent replaces manual expert inspection with automated machine learning systems that can process large volumes of data rapidly. The models learn from training data and automatically identify erroneous triples, eliminating the time-consuming manual review process while maintaining high accuracy through probabilistic reasoning.
4Productivity
If automated data extraction methods are used, then knowledge graphs can be populated, but erroneous data is introduced due to mistakes
Solution Approach 1:
The patent implements a feedback mechanism where the fact-checking system evaluates extracted triples and provides information about their reliability. This feedback loop allows the system to identify and correct erroneous data introduced during extraction, improving overall data accuracy while maintaining extraction productivity.
Data Source
AI summary
A computer-implemented method and system for assigning at least one query triplet to at least one respective class. The at least one respective class is true or false. The method includes the steps of providing the at least one query triplet and a knowledge graph with a plurality of triples and extracting at least one affirmative argument using reinforcement learning on the basis of the at least one query triplet and the knowledge graph. The at least one affirmative argument indicates that the at least one query triplet is true. The method further includes extracting at least one opposing argument using reinforcement learning on the basis of the at least one query triplet and the knowledge graph. The at least one opposing argument indicates that the at least one query triplet is false. The method further includes assigning the at least one query triplet to the at least one respective class using supervised machine learning depending on the at least two arguments.