Machine Learning Parsing of Technical Data Packages for PLM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Technical data packages (TDPs) are difficult to ingest into product lifecycle management (PLM) tools due to challenges in file reading, understanding, and formatting, requiring significant time and manual effort to process and format data for input.
Innovation Solution
Training machine learning models to parse information from multiple files within TDPs, allowing asynchronous processing to generate a consolidated file for PLM tools, which includes identifying parts, components, and instructions with confidence scores, and uploading the processed data into a database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual processing methods are used to ingest TDP data into PLM tools, then data can be processed and understood, but the process requires significant time and engineer effort
Solution Approach 1:
The patent replaces manual mechanical processing methods with machine learning models that automatically parse, understand, and extract data from TDP files. The ML models substitute human engineer analysis with automated computational processes, maintaining data processing reliability while dramatically reducing the time and effort required.
Solution Approach 2:
The patent introduces machine learning models as an intermediary layer between the raw TDP data and the PLM tool. These models act as mediators that translate complex, unstructured TDP contents into structured formats that PLM tools can automatically process, eliminating the need for manual intervention while ensuring accurate data interpretation.
2Productivity
If computer systems attempt to directly read and understand TDP files without specialized processing, then processing speed may be faster, but the systems cannot correctly interpret the data
Solution Approach 1:
The patent applies preliminary action by training machine learning models beforehand to understand the complex structures and semantics of TDP files. These pre-trained models are then deployed to automatically parse and interpret TDP data during ingestion, enabling both fast processing and accurate interpretation without requiring real-time human intervention.
Solution Approach 2:
The patent replaces the inadequate direct reading approach with sophisticated machine learning-based interpretation systems. The ML models substitute simple file reading operations with intelligent analysis capabilities that can understand and extract meaningful information from complex TDP structures, achieving both speed and accuracy.
3Ease of operation
If traditional file processing methods are used to handle multiple TDP files, then the process can be completed, but the complexity of managing multiple file formats and structures increases
Solution Approach 1:
The patent applies universality by designing machine learning models with multi-functional capabilities to handle various TDP file formats, structures, and content types through a single unified processing framework. This universal approach simplifies the user experience while managing the underlying complexity within the ML model architecture rather than requiring multiple specialized processing systems.
Data Source
AI summary
The present disclosure describes analyzing a plurality of files to generate a file containing pertinent information from each of the plurality of files. The plurality of files may be part of a technical data package. The present disclosure describes one or more machine learning algorithms configured to extract parts, components, elements, features, instructions, and the like from text, tables, images, and metadata contained in each of the plurality of files. The parts, components, elements, features, instructions, and the like may be aggregated into a single file, which may be used to generate a bill of materials and other items with a product lifecycle management (PLM).