Patent Search Assistant Using Synonym Expansion and Citation Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing patent search methods are inefficient due to the large volume of data and varying nomenclature, leading to unwieldy results and missed relevant records, especially when users are unfamiliar with alternative terminology.
Innovation Solution
A search assistant method that utilizes keyword extraction, synonym identification, and relevance ranking to highlight and map citations in patent documents, incorporating artificial intelligence for database improvement and user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If keyword-based search is used in patent databases, then search speed is improved, but search accuracy deteriorates due to varying nomenclature and synonyms
Solution Approach 1:
The patent introduces an intermediary processing layer between the user's search query and the patent database. This layer automatically expands keywords by identifying and adding synonyms from a thesaurus, and by extracting related terms from the patent document itself (title, abstract, claims). This intermediary process resolves the contradiction by maintaining fast keyword-based search while improving accuracy through automatic term expansion without requiring manual user input.
Solution Approach 2:
The system performs preliminary actions by pre-identifying and storing synonyms and related terms in a thesaurus database before the actual search occurs. During search, these pre-prepared synonym lists are automatically applied to expand the search query. This preliminary preparation allows the system to quickly handle nomenclature variations without slowing down the search process, resolving the speed-accuracy tradeoff.
2Measurement precision
If users manually narrow search results by class or keyword, then search precision is improved, but user time and effort increase
Solution Approach 1:
The patent implements self-service by enabling the search system to automatically perform the narrowing and refinement functions that would otherwise require manual user effort. The system autonomously expands search terms using synonyms and related terms from the patent document, automatically filters and ranks results by relevance, and presents refined results without requiring the user to manually iterate through multiple search queries. This resolves the contradiction by maintaining high search precision while eliminating the time and effort users would otherwise spend on manual refinement.
3Reliability
If comprehensive keyword search is performed, then recall is improved, but result manageability deteriorates due to overwhelming number of citations
Solution Approach 1:
The patent applies partial action by selectively expanding only the most relevant keywords and synonyms rather than exhaustively searching all possible terms. The system identifies key terms from the patent document (title, abstract, claims) and expands those specifically, rather than performing a blanket expansion of all terms. This selective approach maintains high recall by capturing relevant documents while avoiding the overwhelming proliferation of unrelated citations, thus preserving result manageability.
4Reliability
If users are unfamiliar with alternative terminology, then search completeness deteriorates, but the system should maintain ease of use
Solution Approach 1:
The system performs self-service by automatically compensating for the user's lack of familiarity with alternative terminology. Instead of requiring the user to know and input multiple synonyms or alternative terms, the system autonomously identifies the user's intended search terms from the patent document and automatically expands them with relevant synonyms and related terms from the thesaurus. This resolves the contradiction by maintaining search completeness (finding all relevant documents regardless of terminology variations) while preserving ease of use (the user simply inputs their natural language query without needing to know alternative terminology).
Data Source
AI summary
A search assistant method, the method of searching for patent documents material to elements of a target patent document including: receiving the target patent document including the elements, the target patent document being machine-readable; selecting keywords from the elements; identifying synonyms corresponding to the keywords in a synonym database; forming the keywords and the identified synonyms into a search string; identifying patent documents in a patent database based on the search string; highlighting the keywords and the identified synonyms in the identified patent documents; ranking the identified patent documents based on relevance of the highlighted keywords and the highlighted identified synonyms in each of the identified patent documents to the search string; determining citations in at least one of the ranked identified patent documents for each of the highlighted keywords and the highlighted identified synonyms; and mapping the elements of the target patent document to the determined citations.