Iterative QA Answer Extraction with Context Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional question-answering (QA) systems lack the provision of supporting facts and evidence, resulting in answers with insufficient context, making them difficult for users to understand and process.
Innovation Solution
The technique involves iterative answer and supplemental information extraction using artificial intelligence-based QA systems, where answers are generated and supplemented through iterative processing of user queries and context information, employing supervised and unsupervised methods to extract relevant information without relying on structured databases or knowledge graphs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional QA systems generate answers directly without iterative processing, then the response time is fast, but the answers lack supporting facts and evidence, making them insufficient for user understanding
Solution Approach 1:
The patent segments the answer generation process into multiple iterative steps: initial answer generation, keyword extraction, query modification, supplemental information retrieval, and answer enrichment. Each iteration focuses on extracting specific supplemental information related to keywords from the previous answer, progressively building context-rich responses without requiring the system to process all information simultaneously.
Solution Approach 2:
The system performs preliminary keyword extraction from the initial answer before retrieving supplemental information. By identifying key terms in advance and modifying the query based on these keywords, the system prepares targeted retrieval queries that efficiently locate supporting facts and evidence, reducing the time needed for information gathering while ensuring comprehensive context coverage.
2Loss of information
If the system processes all context information to ensure comprehensive answers, then the answer quality improves, but the system complexity increases
Solution Approach 1:
The patent implements a dynamic iterative process where the system adapts its query based on previous answers. Each iteration modifies the query by incorporating keywords from the previous answer, allowing the system to dynamically adjust its information retrieval strategy. This dynamic approach ensures comprehensive context coverage while maintaining manageable system complexity through adaptive, rather than static, information processing.
3Ease of operation
If conventional QA systems provide direct answers without supplemental information, then the system operation is simple, but the answers are difficult for users to understand and process
Solution Approach 1:
The system employs feedback mechanisms where each retrieved supplemental information is processed and integrated into the final answer. The iterative query modification based on extracted keywords creates a feedback loop that continuously refines the answer quality. This feedback-driven approach ensures that supplemental information directly enhances user understanding while maintaining systematic processing that prevents overwhelming complexity.
Data Source
AI summary
Methods, systems, and computer program products for iterative answer and supplemental information extraction for machine reading comprehension are provided herein. A computer-implemented method includes obtaining a user query and items of context information related to the user query; identifying at least one of the items of context information to be used in connection with generating supplemental information by processing the user query and at least a portion of the multiple items of context information; generating the answer using an artificial intelligence-based question answering system; generating a modified version of the user query based on keywords derived from the generated answer and keywords derived from the user query; generating items of supplemental information for the generated answer by processing the modified version of the user query and the at least one identified item of context information; and performing automated actions based on the items of supplemental information.