Fact Corroboration via Attribute-Value Pair Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines and fact databases face challenges in providing accurate and reliable answers to factual questions due to reliance on single sources, potential for contradictory information, and inclusion of untrustworthy data, which diminishes the quality of the database.
Innovation Solution
A method for corroborating facts by identifying attribute-value pairs associated with a common subject, applying predefined corroboration requirements to filter and validate data, and performing actions on the status of these pairs within a fact repository, utilizing a network of document hosts and a fact repository engine to extract, process, and store factual information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple sources are used to expand the universe of sources, then the types of questions that can be answered increases, but the possibility of contradictory or ambiguous answers increases
Solution Approach 1:
The patent segments the fact verification process into distinct components: identifying attribute-value pairs from multiple sources, evaluating their credibility independently, and then synthesizing results. This segmentation allows the system to handle multiple sources systematically while maintaining reliability through structured evaluation of each source's contribution.
Solution Approach 2:
The patent introduces an intermediary credibility evaluation mechanism that mediates between multiple sources and the final factual answer. This intermediary layer assesses the trustworthiness of each source and attribute-value pair, resolving contradictions by weighting information from more credible sources higher than less credible ones.
2Quantity of substance
If the set of sources is expanded, then more factual information becomes available, but the inclusion of untrustworthy sources increases
Solution Approach 1:
The patent applies local quality by evaluating the credibility of each source and attribute-value pair individually rather than treating all sources uniformly. Different sources receive different credibility assessments based on their specific characteristics, allowing the system to incorporate diverse information while filtering out untrustworthy content through localized credibility evaluation.
3Reliability
If corroboration requirements are applied to filter and validate data, then the reliability of stored facts increases, but the complexity of the data processing increases
Solution Approach 1:
The patent implements preliminary action by establishing credibility evaluation criteria and corroboration requirements before the actual data storage process. This preliminary framework is set up once and then applied systematically to incoming data, reducing the complexity of real-time processing while maintaining high reliability standards through pre-defined evaluation rules.
Data Source
AI summary
A method and system for corroborating facts extracted as attribute-value pairs from multiple sources. Facts associated with a common subject are identified and extracted from multiple sources. From the extracted facts, attributes satisfying a first corroboration requirement are identified. Attribute-value pairs satisfying a second corroboration requirement and having any of the identified attributes that satisfy the first corroboration requirement are identified. Predefined actions may be performed with respect to the identified attribute-value pairs and/or with respect to their status in a fact database.