Synthesizing Electronic Documents for Complex System Fault Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computer-based searching solutions are inadequate for identifying relevant solutions to complex system faults, such as those in aircraft, due to the difficulty in transforming natural language customer reports into effective search queries, leading to manual and resource-intensive processes that require subject matter expertise.
Innovation Solution
A method and system that synthesize electronic documents by filtering out irrelevant terms, using a data dictionary and machine learning models to identify relevant keywords, and weighting terms based on their importance, allowing for autonomous identification of similar issues and solutions within a data repository.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If conventional computer-based searching solutions are used to identify relevant solutions to complex system faults, then the search process can be automated to some extent, but the ability to accurately understand and transform natural language customer reports into effective search queries remains insufficient, requiring manual intervention and subject matter expertise
Solution Approach 1:
The patent introduces an intermediary system comprising natural language processing components, data dictionaries, and machine learning models that mediate between customer reports and the data repository. This intermediary automatically transforms natural language queries into effective search queries by identifying key terms, applying filtering rules, and weighting terms based on relevance, thereby eliminating the need for manual subject matter expert intervention while maintaining high accuracy in query transformation
2Ease of operation
If manual processes with subject matter expertise are used to identify relevant solutions, then accuracy in understanding natural language reports can be maintained, but resource consumption and time requirements increase significantly
Solution Approach 1:
The system enables self-service by automatically processing customer reports without requiring manual subject matter expert intervention. The natural language processing system, data dictionaries, and machine learning models work autonomously to understand natural language reports, identify relevant terms, and retrieve appropriate solutions from the data repository, thereby maintaining high accuracy while dramatically improving resource efficiency and reducing time requirements
Solution Approach 2:
The patent replaces the mechanical system of manual subject matter expert analysis with an automated information processing system. Machine learning models, natural language processing algorithms, and computational weighting mechanisms substitute for human cognitive processes, enabling the system to understand and analyze natural language reports with high accuracy while consuming significantly fewer human resources and time
3Reliability
If comprehensive text analysis is performed on all electronic documents to identify relevant solutions, then completeness of information retrieval can be ensured, but processing time and computational resources increase
Solution Approach 1:
The patent extracts only the most relevant information from electronic documents by applying filtering rules that identify and remove unnecessary portions of text. The system extracts key terms, phrases, and concepts that are most likely to be relevant to the customer report, rather than analyzing every word in every document. This extraction process, combined with term weighting based on relevance, ensures that the most important information is retrieved quickly while minimizing processing time and computational resources
4Quantity of substance
If data repositories grow larger to include more historical solutions and information, then the completeness and value of available solutions increase, but the difficulty of identifying relevant information among the growing volume of data increases
Solution Approach 1:
The patent replaces manual information detection with automated machine learning models and natural language processing systems. These computational systems can efficiently search through and analyze large volumes of data in the repository, identifying relevant information based on learned patterns and relationships. The system automatically weights terms, applies filtering rules, and ranks results by relevance, making the process of identifying relevant information scalable to large data volumes without proportionally increasing difficulty
Solution Approach 2:
The patent introduces an intermediary information processing layer that mediates between the large data repository and the customer report. This intermediary system uses machine learning models, data dictionaries, and automated query transformation to navigate the large volume of available solutions and identify the most relevant ones. The intermediary translates natural language queries into effective search queries that can efficiently retrieve relevant information from the extensive repository, thereby managing the complexity of large data volumes
Data Source
AI summary
Techniques for identifying relevant, natural language documents within a data repository. An electronic document specifying natural language text describing an issue with a complex system is received. One or more portions are removed from the electronic document. The portions are determined to satisfy predefined filtering rules. A first set of terms are determined using a data dictionary structure, and a second set of terms are determined, where at least one term in the second set of terms satisfies at least one predefined pattern matching rule. A third set of terms are determined by processing the electronic document as an input to a machine learning model trained to recognize relevant terms within the electronic document. A synthesized electronic document is generated from the first, second and third sets of terms, and the synthesized electronic document is used to identify a set of relevant documents within the data repository.