Ambiguity Screening in AI Question Generation for Multiple Answers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing question generation systems fail to accurately generate questions corresponding to multiple answers, leading to ambiguity and inaccuracy.
Innovation Solution
A question generation apparatus and method that includes an acquiring process to gather character strings and associated texts, a question generating process to create questions for each string, and an ambiguity judging process to identify and output questions with multiple answers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a question generator generates a single question from a single answer, then the question generation process is simple, but the generated question is not necessarily appropriate for multiple answers
Solution Approach 1:
The system segments the question generation process into multiple independent stages: (1) generating individual questions for each answer, (2) evaluating question-answer pairs, (3) selecting questions that work with multiple answers. This segmentation allows each stage to be optimized independently, resolving the contradiction between simplicity and accuracy.
Solution Approach 2:
The system performs preliminary actions by first generating questions for individual answers before evaluating their suitability for multiple answers. This preliminary generation followed by evaluation allows the system to maintain simple generation processes while ensuring final question accuracy through subsequent filtering and selection.
2Adaptability or versatility
If multiple questions are generated for each answer, then the coverage of possible answers increases, but the complexity of question generation increases
Solution Approach 1:
The system creates questions that serve multiple functions simultaneously - each generated question is designed to work with multiple different answers, making the question universal rather than answer-specific. This multi-functionality approach increases coverage without proportionally increasing system complexity.
Solution Approach 2:
The system implements feedback mechanisms where generated questions are evaluated against multiple answers, and this evaluation feedback is used to select and refine questions. The feedback loop ensures high coverage by continuously verifying question suitability, while the automated nature of the feedback process manages system complexity.
3Productivity
If question generation is performed without evaluating ambiguity, then the generation speed is fast, but the accuracy of generating questions with multiple answers is low
Solution Approach 1:
The system performs preliminary question generation without ambiguity evaluation to maintain high speed, then conducts a separate evaluation phase to ensure accuracy. This preliminary action followed by verification allows the system to achieve both fast generation and high accuracy by separating the speed-critical and accuracy-critical stages.
Solution Approach 2:
The system segments the process into a fast generation phase that produces multiple candidate questions, and a separate evaluation phase that filters for accuracy. This segmentation allows each phase to be optimized for its specific goal - speed for generation, accuracy for evaluation - resolving the contradiction between productivity and precision.
Data Source
AI summary
At least one processor included in a question generation apparatus carries out: an acquiring process of acquiring a plurality of character strings and pieces of text; a question generating process of generating, for each of the plurality of character strings, a question an answer to which is that character string; an ambiguity judging process of judging, for each of a plurality of questions, whether there are a plurality of answers to that question; and an outputting process of outputting a question which has been judged to be the question to which there are a plurality of answers.