Scenario Passage Pair Recognizer for Coherent Causality Chaining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for automatically generating scenarios by chaining causalities often produce inconsistent scenarios due to insufficient comprehension of context, leading to erroneous outcomes like "swallows barium→go through an X-ray examination→board on a plane," which results from chaining causalities across different contexts without proper attention to their respective contexts.
Innovation Solution
A scenario passage pair recognizer and classifier system that uses machine learning to assess the reliability of scenario candidates by searching for supporting passages in documents, extracting features, and calculating scores to determine coherence and plausibility, ensuring consistent context across chained causalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If causalities are chained by matching effect and cause portions without context filtering, then scenario generation productivity increases, but scenario reliability deteriorates due to inconsistent contexts
Solution Approach 1:
The patent applies preliminary action by performing context verification before final scenario generation. The system pre-processes causalities by extracting and comparing context information (such as event types, entities, and relationships) to ensure consistency before chaining them into scenarios. This preliminary context matching prevents inconsistent scenarios from being generated in the first place, resolving the contradiction between high productivity and high reliability.
2Device complexity
If a simple word overlap filter is used to assess causality consistency, then device complexity is reduced, but measurement precision of context consistency deteriorates
Solution Approach 1:
The patent applies parameter changes by transforming the context consistency assessment from simple word overlap measurement to multi-dimensional parameter comparison. The system evaluates context consistency using multiple parameters including event type similarity, entity relationship compatibility, temporal sequence alignment, and semantic coherence. This multi-parameter approach significantly improves measurement precision while maintaining manageable system complexity through efficient algorithms.
Data Source
AI summary
A scenario passage pair recognizer includes: a text passage searching unit searching a set of text passages each including no more than a certain number of sentences of a document, and within which all noun phrases included in a scenario candidate co-occur; a feature extracting unit extracting a feature from each combination of the scenario candidate and each searched support passage; a classifier outputting a score indicating reliability of the scenario candidate based on the support passage as a source of the feature; and a score accumulating unit and a maximum value selecting unit, accumulating the scores output from the classifier and selecting the maximum value as the reliability of the scenario candidate. The scenario classifier determines plausibility of the scenario candidate as a causality based on the feature including the score output from the scenario passage pair recognizer.