Automated Information Retrieval Using Combination Term Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated information retrieval systems using keyword matching often fail to provide semantically relevant search results, leading to inefficiencies in processing and computational costs.
Innovation Solution
The system generates a combination term vector by creating word vectors and applying term vector rules such as elementwise minimum, maximum, or weighted mean values, which are then normalized and combined to represent phrases or sentences, enabling more accurate and efficient semantic searches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If keyword matching techniques are used for search and content discovery, then the system can process search queries, but the search results do not provide semantically relevant results and accuracy deteriorates
Solution Approach 1:
The patent transforms the search parameter from discrete keyword matching to continuous semantic vector representation. By representing both search queries and content as vectors in a semantic space, the system changes the fundamental parameter of comparison from exact string matching to geometric distance measurement, enabling semantically equivalent terms to be identified even when keywords differ.
Solution Approach 2:
The patent replaces the mechanical keyword matching system with a mathematical vector-based semantic analysis system. Instead of mechanically comparing strings character by character, the system uses vector operations (dot products, cosine similarity) to measure semantic relationships, substituting a rigid mechanical process with a flexible mathematical model that captures meaning.
2Productivity
If automated information retrieval systems process large volume of natural language data, then comprehensive search coverage is achieved, but processing time increases and speed deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing semantic vectors for all content in the database during an offline preprocessing phase. This allows the system to avoid performing complex semantic analysis during online search operations, as the vectors are already prepared and stored, enabling fast retrieval through simple vector comparison operations.
3Speed
If keyword matching is used to provide search results, then the system can return results quickly, but computational accuracy and semantic relevance deteriorate
Solution Approach 1:
The patent creates a simplified copy or representation of the semantic meaning in vector form, which can be quickly compared and manipulated. Instead of performing complex natural language understanding during search, the system uses pre-computed vector copies that capture the essential semantic information, enabling fast and accurate similarity measurements through efficient vector operations.
Data Source
AI summary
An automated information retrieval system includes a processor coupled to a communications interface and a non-transitory memory. The processor is configured to read instructions from the non-transitory memory to cause the system to perform operations comprising receiving a term including a plurality of words and generating a plurality of word vectors corresponding to the plurality of words respectively. A plurality of term vectors associated with the term is generated using the plurality of word vectors based on a plurality of term vector rules respectively. A combination term vector associated with the term is generated by combining the plurality of term vectors. One or more search results from a database for the term are provided using the combination term vector.