Entity Matching Using Visual Information Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional entity matching methods fail to effectively link records from multiple databases due to the lack of consideration for visual information, leading to inaccuracies in identifying matching entities.
Innovation Solution
A method and system that utilize visual information by encoding and classifying entities using a combination of text, numeric, and visual encoders, with a classifier determining whether entities match or not based on the encoded features, incorporating visual embeddings generated by convolutional neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional entity matching methods are used, then the process is simple and fast, but the accuracy of identifying matching entities is poor
Solution Approach 1:
The patent combines multiple encoding approaches (text encoder, numeric encoder, visual encoder) into a unified entity matching system. The encoders process different types of data (text, numbers, images) and their outputs are integrated through a classifier to determine entity matches, achieving higher accuracy by merging diverse data sources
Solution Approach 2:
The system creates a universal encoder framework that can handle multiple data types (text, numeric, visual) through separate encoder modules. This multi-functional approach allows the same matching system to process diverse entity attributes, improving accuracy without requiring separate specialized systems for each data type
2Measurement precision
If visual information is incorporated into entity matching, then the accuracy improves, but the processing time and computational resources increase
Solution Approach 1:
The patent segments the encoding process into separate modules (text encoder, numeric encoder, visual encoder), each handling specific data types independently. This segmentation allows parallel processing of different data types and enables the system to process only relevant visual information when needed, reducing unnecessary computational overhead while maintaining high accuracy
Data Source
AI summary
One embodiment provides a method comprising receiving a first set of information corresponding to a first entity and receiving a second set of information corresponding to a second entity. Each set of information received corresponding to each entity comprises visual information relating to the entity. The method further comprises encoding each set of information, and classifying the first entity and the second entity as either matching entities or non-matching entities based on the encoding.