Multi-Segment Text Search for Precise Patent Claim Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional search engines require users to input specific keywords, which can be challenging, and are not effective for finding content similar to an entire block of text, particularly in the context of patent claims.
Innovation Solution
A computer-implemented system splits textual inputs, such as patent claims, into segments and uses a text similarity algorithm to identify similar segments from stored documents, enabling precise identification of portions that invalidate or render claims unpatentable due to lack of novelty or obviousness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional keyword-based search is used, then the search engine can operate with simple algorithms, but the user must determine the right keywords and the system cannot effectively find content similar to an entire block of text
Solution Approach 1:
The patent segments the input text block into multiple text segments and processes each segment independently through the search engine. This segmentation allows the system to handle large blocks of text by breaking them into manageable units, enabling both keyword-based and similarity-based search operations on smaller, more manageable segments while maintaining the ability to find relevant content throughout the entire input block
Solution Approach 2:
The patent introduces text embeddings as an intermediary representation between the input text segments and the search engine. These embeddings transform textual data into numerical vectors that capture semantic meaning, allowing the search engine to perform similarity comparisons without requiring users to manually select keywords. The embeddings serve as a bridge that enables automatic content similarity detection
2Measurement precision
If the system processes entire text blocks for similarity search, then it can find comprehensive similar content, but the processing time and computational resources increase significantly
Solution Approach 1:
The patent divides the input text block into multiple smaller text segments before processing. This segmentation reduces the computational complexity of similarity search by comparing smaller units rather than entire large text blocks. The system can process multiple segments in parallel and aggregate results, maintaining comprehensive similarity detection while significantly reducing processing time and resource requirements
Solution Approach 2:
The patent processes text segments independently and selectively, rather than processing the entire text block as a single unit. This partial action approach allows the system to identify and focus on the most relevant segments for similarity matching, reducing overall processing time while maintaining accuracy in identifying similar content
3Productivity
If manual keyword selection is required, then the search can be targeted to specific terms, but the user expertise and time investment increase
Solution Approach 1:
The patent implements a self-service search system where the text embedding model automatically extracts semantic features and performs similarity search without requiring user input of keywords. The system serves itself by automatically transforming input text segments into embeddings and comparing them against the database, eliminating the need for users to manually select or understand appropriate search terms while maintaining high search efficiency
Data Source
AI summary
Systems and methods may be provided for performing a search on an input text block. The input text block may be split into a plurality of input text segments. A text similarity algorithm may be used to find similar stored text segments to each of the plurality of input text segments.