Document Search System Using Parallel AND-OR Tree Sub-Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Trademark document clearance searches are inefficient due to combinatorial explosions of queries, redundant results, and lack of term proximity information in search engines, making it difficult to accurately and quickly find relevant documents.
Innovation Solution
A computer-implemented document search system constructs an AND-OR or AND-OR_N tree to logically combine terms from two lists, processes sub-queries in parallel, and returns results with ordinal positions of matching terms, eliminating duplicates and improving query efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If trademark document clearance searches use combinatorial queries with hundreds of terms, then search coverage is improved, but query processing time increases significantly
Solution Approach 1:
The patent divides the large query into multiple sub-queries by selecting terms from different lists (trademark terms and industry terms) and processing them in parallel. This segmentation allows the search system to handle complex queries efficiently by breaking them into manageable units that can be executed simultaneously, reducing overall processing time while maintaining comprehensive search coverage.
Solution Approach 2:
The patent implements dynamic query processing where the search system adapts the query structure based on the input lists. By using AND-OR logic trees that dynamically combine terms from trademark lists and industry term lists, the system can efficiently process variable query complexities without fixed processing time penalties.
2Reliability
If search engines submit multiple separate queries for trademark combinations, then search completeness is improved, but redundant results increase
Solution Approach 1:
The patent merges multiple query results by using set operations on the results from different sub-queries. Instead of treating each query independently and risking redundancy, the system combines results from trademark term combinations with industry term combinations using logical operations that eliminate duplicate documents, ensuring completeness while reducing redundancy.
Solution Approach 2:
The patent introduces an intermediary result processing layer that receives results from multiple sub-queries and mediates between them through deduplication logic. This intermediary layer ensures that documents appearing in multiple query results are properly consolidated, maintaining search completeness while eliminating redundant entries in the final output.
3Reliability
If search results include all matching documents, then search thoroughness is improved, but result quality decreases due to lack of focus
Solution Approach 1:
The patent applies local quality by providing different levels of detail in the search results based on the query structure. Results are enhanced with metadata about which specific trademark terms and industry terms matched, allowing users to quickly identify relevant results without seeing all possible matches. This localized enhancement of result quality maintains thoroughness while improving usability.
Data Source
AI summary
A computer-implemented search method and system is provided whereby a file is received including a query, typically for trademark search, having a first list of query terms and a second list of industry terms. The first and second lists may include tens or hundreds of terms. An AND-OR tree is constructed to represent the query terms and the industry terms and traversed to logically combine each of the query terms with each of the industry terms to form sub-queries which are executed in parallel. The results from each sub-query execution are limited according to a sub-query result limit. The sub-query results are combined to construct a query result for the original query included in the file. Duplicate results are removed from the query result and an ordinal number is included for each term in the query result before returning the query result.