Sentence Structure Conversion for Hypothesis Support Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing text analysis engines, such as IBM Debater Pro/Con API, often produce inaccurate scores for the similarity between a hypothesis and an opinion due to being trained on unrelated corpora, requiring a more accurate method to determine support for a hypothesis from opinion sentences.
Innovation Solution
A system and method that converts opinion sentences using various sentence structure conversion methods, calculates the difference between manually defined proximity labels and initial scores, and adopts conversions that improve the score proximity, creating sub-opinions and calculating final support scores using a processor-based text analysis engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sentence structure conversion methods are applied to opinion sentences, then the accuracy of hypothesis support determination is improved, but the complexity of the processing system increases
Solution Approach 1:
The patent segments the opinion sentence processing into multiple conversion methods (deletion, replacement, splitting) that can be applied independently. Each conversion method processes specific portions of the opinion sentence, allowing the system to handle complex sentences through modular operations rather than requiring a monolithic complex processor.
Solution Approach 2:
The patent applies sentence structure conversion methods as preliminary actions before the main scoring process. By pre-converting opinion sentences into multiple variants (deleted portions, replaced portions, split portions), the system prepares optimized inputs that will yield more accurate hypothesis support scores without requiring complex real-time processing during scoring.
2Measurement precision
If multiple sentence structure conversion methods are used, then the score accuracy closer to manual labels is achieved, but the processing time increases
Solution Approach 1:
The patent applies partial action by selectively using conversion methods based on the specific opinion sentence characteristics. Not all conversion methods are applied to every sentence - instead, the system chooses appropriate conversions (deletion, replacement, or splitting) based on the sentence structure and content, achieving good accuracy without the overhead of applying all possible conversions universally.
Solution Approach 2:
The patent incorporates feedback mechanisms where the scores from converted opinion sentences are compared against manual proximity labels. This feedback loop allows the system to identify which conversion methods produce scores closest to manual labels, enabling the system to learn and prioritize effective conversion strategies, thereby reducing processing time for future similar cases.
3Reliability
If opinion sentences are converted and subdivided into sub-opinions, then the reliability of similarity assessment is improved, but the quantity of processing operations increases
Solution Approach 1:
The patent segments opinion sentences into sub-opinions through the splitting conversion method. When an opinion sentence contains multiple distinct claims or arguments, it is divided into separate sub-opinions that can be evaluated independently against the hypothesis. This segmentation improves reliability by ensuring each distinct argument is properly assessed, while the modular nature of segmentation allows efficient processing through systematic division rather than exhaustive analysis.
Solution Approach 2:
The patent extracts relevant portions of opinion sentences through the deletion conversion method. By removing irrelevant or redundant portions of opinion sentences, the system focuses processing resources on the core arguments that directly support or refute the hypothesis. This extraction improves reliability by eliminating noise while reducing the quantity of processing operations needed.
Data Source
AI summary
A method is provided for determining support of a hypothesis by opinion sentences. The method converts sentence structures in the opinion sentences using various sentence structure conversion methods to obtain converted opinion sentences. For each converted opinion sentence, the method calculates a difference between a proximity label value indicating proximity to the hypothesis and an intermediate score before and after a conversion, adopts the conversion responsive to a condition being met relative to the difference, and adopts the opinion sentence instead responsive to the condition being unmet. The method creates sub-opinions using the various methods applied to adopted conversions and opinion sentences, and obtains an intermediate score for each sub-opinion. The method represents an amount of support for the hypothesis by obtaining and displaying a final score for each adopted conversions and opinion sentences based on the intermediate scores for the sub-opinions and for adopted conversions and opinion sentences.