Multi-Modal Intelligence Data Fusion via Entity Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for fusing intelligence data from multiple modalities face challenges in entity resolution, disambiguation, and collapsing links into meaningful relationships, leading to complex networks with reduced exploitation effectiveness.
Innovation Solution
A method and system for cross-INT entity resolution that optimizes the mapping of identifiers to entities by considering compatibility of attributes, mutual information across interaction data sources, and fit with behavior models, using a multi-term objective function to minimize network complexity and maximize exploitation effectiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple intelligence modalities are fused to improve entity resolution accuracy, then entity fusion accuracy is improved, but network complexity increases
Solution Approach 1:
The patent segments the entity resolution process into distinct phases: data collection from multiple modalities, data fusion to create unified entity representations, and network construction from fused entities. This segmentation allows each phase to be optimized independently, improving entity resolution accuracy while managing network complexity through structured processing stages.
Solution Approach 2:
The patent introduces an intermediary data fusion layer that processes and reconciles data from multiple intelligence modalities before constructing the final network. This intermediary layer acts as a mediator that transforms raw multi-modal data into standardized entity representations, improving resolution accuracy while preventing complexity propagation to the network level.
2Productivity
If comprehensive data fusion is performed to enhance exploitation effectiveness, then exploitation effectiveness is improved, but processing complexity increases
Solution Approach 1:
The patent performs preliminary data fusion and entity resolution operations before network construction and analysis. By pre-processing the data to create unified entity representations and resolve ambiguities in advance, the system enhances exploitation effectiveness during network analysis while reducing the processing complexity burden during the main analytical phase.
3Measurement precision
If entity mappings are optimized using multiple criteria to improve mapping accuracy, then mapping accuracy is improved, but computational requirements increase
Solution Approach 1:
The patent applies different optimization criteria and weighting schemes to different aspects of entity mapping: attribute compatibility, link structure consistency, and behavior model fit. Each aspect is optimized with appropriate local quality measures rather than applying a single global optimization criterion, improving overall mapping accuracy while distributing computational requirements across manageable sub-problems.
Data Source
AI summary
Disclosed is a method for fusing interaction data, such as intelligence data, comprising, embodying collections of interaction data from different interaction data sources in interaction graphs, defining a plurality of mappings of identifiers to entities, associating each mapping with a fused interaction graph, and identifying an optimal mapping by evaluation of compatibility of identifier attributes, mutual information across interaction data sources, and/or fit with one or more behavior models. Edges in the fused graph can be collapsed. Also claimed are a computer system and a computer-readable medium for fusing interaction data.