Full-Feature Object Identity Queries Using Multi-Dimensional Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in efficiently querying and comparing entities due to non-measurable and non-comparable data formats, making it difficult to accurately search and identify entities across various domains.
Innovation Solution
A method and system that utilize multi-dimensional feature vectors to represent entities, allowing for queries to be converted into query vectors for comparison with these feature vectors, enabling accurate matching and identification of entities based on their attributes, with dynamic updates and normalization to ensure relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If entities are represented using non-measurable and non-comparable data formats, then data can be stored in diverse formats from various sources, but querying and comparing entities becomes difficult and inefficient
Solution Approach 1:
The patent introduces multi-dimensional feature vectors as an intermediary representation layer between diverse entity data formats and query operations. These feature vectors serve as a common language that translates various data formats (text, numerical, categorical) into a unified mathematical representation, enabling efficient comparison and querying while preserving the versatility of handling different data types from multiple sources
2Productivity
If multi-dimensional feature vectors are used to represent entities, then querying efficiency and matching accuracy are improved, but system complexity increases
Solution Approach 1:
The patent segments entity representation into multiple independent feature vectors, where each vector corresponds to a specific attribute or aspect of the entity. This segmentation allows the system to handle complex entities by breaking them down into manageable components, each can be processed and compared independently, reducing the cognitive load and simplifying the overall querying mechanism despite the increased dimensional complexity
3Measurement precision
If feature vectors are dynamically updated and normalized, then matching accuracy is improved, but computing resources and processing time increase
Solution Approach 1:
The patent implements preliminary normalization and standardization of feature vectors during the data ingestion and storage phase, rather than performing these operations repeatedly during query processing. This preliminary action prepares the data in advance, transforming it into a ready-to-compare format that reduces computational overhead during actual querying operations while maintaining high matching accuracy
Data Source
AI summary
A method of querying entities according to their identifying attributes, comprising accessing a repository storing a plurality of multi-dimensional feature vectors each uniquely representing a respective one of a plurality of entities and comprising a plurality of feature vectors each comprising a plurality of features relating to a respective one of a plurality of attributes of the respective entity, receiving one or more queries each defining a search for one or more of the plurality of entities, converting each query to one or more query vectors each comprising features relating to one or more search attributes defined by the respective query, comparing between the query vector(s) feature vectors of at least some of the multi-dimensional feature vectors, and outputting a response to the query(s) listing each entity represented by a multi-dimensional feature vector comprising one or more feature vectors estimated to match at least partially the query vector(s).