Unstructured Contract Data Structuring via Segmentation and Model Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The field of machine reading comprehension faces challenges in processing unstructured contract and legal data due to its vastness, variability in formats, and the complexity of legal language, which makes it difficult for algorithms to parse and analyze effectively, leading to impractical and error-prone risk management and analysis.
Innovation Solution
A computer-implemented method transforms unstructured data, such as a PDF image of a contract document, into a structured set of data by segmenting, classifying, and grouping data blocks, and extracting key terms using separately trained machine learning models, allowing for easier navigation and analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine learning and artificial intelligence techniques are used to process contract and legal data, then automation and efficiency are improved, but accuracy and reliability deteriorate due to the wide variety of formats, styles, and legal language that algorithms struggle to parse
Solution Approach 1:
The patent segments unstructured contract documents into structured data blocks with specific attributes (e.g., party names, dates, obligations). This segmentation allows the system to process different portions of contracts using specialized algorithms tailored to specific document sections, improving both automation efficiency and parsing accuracy simultaneously
Solution Approach 2:
The patent introduces an intermediary structured data format that sits between the unstructured contract document and the analysis algorithms. This intermediary structure standardizes the representation of contract elements, enabling algorithms to process diverse contract formats reliably while maintaining high automation efficiency
2Ease of operation
If templates and standardized clauses are used to structure contract data, then data organization and analysis are improved, but flexibility and adaptability worsen due to the large amount of wild texts that have been modified from templates
Solution Approach 1:
The patent applies local quality by treating different portions of contract documents with different levels of structuring. Standardized clauses are heavily structured for easy analysis, while modified wild texts are structured with sufficient detail to maintain adaptability. This allows the system to achieve both data organization and flexibility simultaneously
3Device complexity
If a single extraction model is used for all segments, then device complexity is reduced, but measurement precision and extraction accuracy deteriorate
Solution Approach 1:
The patent implements a dynamic model selection mechanism that automatically chooses the appropriate extraction model based on the characteristics of each segment. This dynamic approach allows the system to maintain high extraction accuracy for different segment types while managing complexity through automated model routing rather than manual configuration
Data Source
AI summary
There is provided a computer implemented method of transforming an unstructured set of data to a structured set of data. In some examples, the method comprises segmenting the unstructured set of data into segments, classifying each segment, extracting key terms from each segment using an extraction model, the extraction model selected from a plurality of extraction models based on the classification of the segment, generating the structured set of data using the segments and the extracted key terms.