Network Analysis Tool for Cryptocurrency Fraud Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting illicit activities in cryptocurrency networks, such as centrality measures, are computationally expensive and discrete, requiring frequent updates, which can be inefficient.
Innovation Solution
A computer-implemented method for analyzing a network of information by accessing a graph dataset, determining initial risk parameters of nodes, and updating these parameters based on a diffusion problem to identify outlier nodes potentially involved in illicit activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If centrality measures (eigenvector centrality or Katz centrality) are used to identify influential nodes, then nodes connected to other high-scoring nodes can be detected indicating potential illicit sub-networks, but the computational cost becomes expensive and frequent updates are required
Solution Approach 1:
The patent changes the parameter from discrete centrality scores to continuous risk parameters that evolve over time according to a diffusion process. This allows the system to capture dynamic risk propagation through the network while avoiding repeated computation of expensive centrality measures. The risk parameter update follows a differential equation that models how risk diffuses continuously across network edges, transforming the discrete update problem into a continuous parameter evolution problem.
Solution Approach 2:
The patent introduces dynamic risk parameters that continuously evolve according to a diffusion process rather than using static or frequently recomputed centrality measures. The system models risk as a dynamic quantity that flows through the network over time, governed by differential equations. This dynamic approach allows the system to adapt to changing network conditions without requiring frequent discrete recalculations, thereby reducing computational complexity while maintaining detection reliability.
2Adaptability or versatility
If discrete algorithms are used for network analysis, then updates can be performed, but frequent updates are required which increases computational expense
Solution Approach 1:
The patent implements continuous risk parameter updates through a diffusion process governed by differential equations, eliminating the need for discrete periodic recalculations. The risk parameters evolve continuously as new information becomes available, with updates occurring naturally through the mathematical model rather than requiring explicit discrete algorithm executions. This continuous action approach maintains adaptability while dramatically improving computational efficiency by avoiding repeated full-network analysis.
Solution Approach 2:
The patent pre-establishes the diffusion model and risk propagation rules before actual analysis begins. The system sets up the differential equations and parameter relationships in advance, allowing risk parameters to be updated efficiently as the process unfolds. This preliminary configuration enables the system to handle updates through simple parameter adjustments rather than requiring complete re-analysis, thereby improving productivity while maintaining adaptability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented method for analysing a network of information comprising a plurality of interconnected nodes, the method comprising: accessing a graph dataset comprising graph data representative of a graph comprising a plurality of nodes and a plurality of edges connecting the nodes; determining initial risk parameters of the nodes of the graph; and updating risk parameters of one or more nodes of the graph based on the initial risk parameters of the nodes of the graph, the updated risk parameters representing a likelihood of the associated nodes being outlier nodes.