Automated Reading Comprehension Entity Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines lack the ability to effectively determine similarities and differences between entities in text segments and assess the likelihood of actions being performable by subjects, which hinders enhanced search results and automated reading comprehension.
Innovation Solution
A computer-implemented method that identifies entities and their attributes in text segments, determines similarities and differences, and evaluates the likelihood of actions being performable by subjects based on reference subjects and actions from a corpus of documents, using techniques such as regular expressions, machine learning, and co-reference resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If search engines provide basic document retrieval and ranking, then search functionality is available, but the ability to determine similarities and differences between entities and assess action performability is lacking
Solution Approach 1:
The patent segments text into structured components (entities, attributes, subjects, actions) and processes each separately. Entity identification, attribute extraction, and action analysis are performed as distinct steps, allowing the system to systematically determine similarities/differences between entities and assess action performability without overwhelming complexity.
Solution Approach 2:
The patent introduces intermediary processing layers including co-reference resolution mechanisms and reference subject-action databases. These intermediaries bridge the gap between raw text and meaningful entity comparisons, enabling the system to infer entity relationships and action performability through intermediate representation and comparison stages.
2Productivity
If automated reading comprehension is implemented without entity comparison capabilities, then basic text processing is possible, but robust information extraction and understanding are limited
Solution Approach 1:
The patent replaces simple keyword-matching mechanisms with sophisticated entity identification and attribute comparison systems. Instead of mechanical text search, the system uses entity extraction, attribute classification, and similarity determination algorithms to achieve more accurate text understanding and reading comprehension.
Solution Approach 2:
The patent performs preliminary entity identification and attribute extraction before conducting reading comprehension analysis. By pre-processing text to identify entities, their attributes, and relationships in advance, the system prepares structured data that enhances subsequent comprehension tasks and improves overall accuracy.
3Reliability
If entity attributes are not compared, then text processing is simpler, but robust information about text segments cannot be provided
Solution Approach 1:
The patent applies local quality by focusing attribute comparison on specific entities and their relevant attributes rather than performing global text analysis. The system identifies entities, extracts their specific attributes, and compares only those attributes that are meaningful for the given context, providing robust information without unnecessary complexity.
Data Source
AI summary
Methods and apparatus are disclosed for determining similarities and/or differences between entities in a segment of text based on various signals are presented, and for determining one or more likelihoods that one or more subjects found in a segment of text are capable of performing one or more associated actions based on various signals.