Automated Passage Mapping for Reading Comprehension Test Items
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for developing test items from text-based content are manual, laborious, time-consuming, subjective, and inconsistent, and lack automated assistance for identifying suitable passages and generating questions, especially due to the vast volume and complexity of content.
Innovation Solution
The use of natural language processing (NLP) and machine learning (ML) models to analyze passages, identify skill-specific textual patterns, and generate test items by highlighting relevant language cues, thereby automating the passage-mapping process and providing a systematic approach for content specialists.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual passage mapping is used by content specialists, then test item development can be performed, but the process becomes time-consuming and inconsistent
Solution Approach 1:
The patent replaces the manual mechanical process of passage mapping with an automated computer-based system using natural language processing and machine learning algorithms. The system automatically analyzes passages, identifies skill-specific textual patterns, and generates test item opportunities without human intervention in the mapping process, thereby eliminating time consumption and ensuring consistency.
Solution Approach 2:
The system enables self-service automated passage mapping where the computer system independently performs the entire passage mapping process without requiring content specialist involvement. The NLP and ML models autonomously identify testable content, patterns, and skill alignments, making the process both faster and more reliable.
2Quantity of substance
If the volume and complexity of pre-existing text-based content increases, then more test item opportunities are available, but selecting suitable portions becomes more difficult
Solution Approach 1:
The system extracts only the relevant testable portions from large volumes of pre-existing content by using NLP to identify skill-specific textual patterns and semantic structures. The ML models filter through the vast content to extract only those portions that meet predefined criteria for test item development, making the selection process systematic and efficient.
Solution Approach 2:
The system changes the parameters of content analysis by using multiple NLP techniques (tokenization, part-of-speech tagging, named entity recognition) and ML models to evaluate content based on various linguistic and semantic parameters. This multi-parameter approach enables the system to navigate and select from vast content volumes by evaluating multiple dimensions simultaneously.
3Productivity
If automated passage mapping is implemented using NLP and ML models, then efficiency and consistency improve, but system complexity increases
Solution Approach 1:
The automated system is segmented into distinct functional modules: NLP processing module for linguistic analysis, ML modeling module for pattern recognition, and test item generation module for output creation. This segmentation allows each component to be developed, maintained, and optimized independently, managing overall system complexity while maintaining high productivity.
Data Source
AI summary
The present disclosure provides systems and methods for providing interface rendering of passages for test item generation. Methods and systems can utilize natural language processing (NLP) and machine learning (ML) models to analyze passages that were provided in historical testing environments (e.g., first passage), determine one or more similar search strings that are also found in a current passage (e.g., second passage), and provide the first and second passages to a user interface of a user device. On the display, the search string in the historical and current passages may be highlighted at the user interface to identify the similarities.