Iterative Blockchain Graph Embedding for Scalable Risk Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for identifying fraudulent transactions in blockchain networks face challenges in scaling to the increasing size and dynamic nature of cryptocurrency transactions, making it difficult to efficiently generate risk scores for blockchain addresses.
Innovation Solution
Implementing a soft binary classification model with a node2vec algorithm for graph-based feature extraction and behavioral-based features to generate risk scores, using incremental graph recomputation and distributed task-based approaches to reduce computational overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing techniques are used to identify fraudulent transactions in blockchain networks, then fraud detection capability is maintained, but the system cannot scale to increasing size and dynamic nature of cryptocurrency transactions
Solution Approach 1:
The patent segments the blockchain transaction analysis into graph-based feature extraction and behavioral-based feature extraction. The graph-based features capture topological relationships between addresses, while behavioral-based features analyze transaction patterns. This segmentation allows the system to handle large transaction volumes efficiently by processing different aspects of fraud detection through specialized, scalable mechanisms.
Solution Approach 2:
The patent implements dynamic fraud detection by continuously updating graph representations and behavioral features as new transactions occur. The system adapts to the dynamic nature of cryptocurrency transactions by recalculating risk scores based on evolving transaction patterns and network topology, enabling scalability without sacrificing detection accuracy.
2Measurement precision
If comprehensive fraud detection analysis is performed on all blockchain transactions, then detection accuracy is improved, but computational overhead increases significantly
Solution Approach 1:
The patent applies partial action by focusing computational resources on high-risk transactions identified through preliminary filtering. Instead of performing comprehensive analysis on all transactions, the system uses graph-based and behavioral features to identify suspicious patterns, then applies more intensive analysis only to transactions that exceed certain risk thresholds, significantly reducing overall computational overhead while maintaining high detection accuracy.
Solution Approach 2:
The patent replaces traditional rule-based fraud detection mechanisms with machine learning models that analyze graph-based and behavioral features. This substitution enables more efficient processing by leveraging patterns learned from historical data, reducing the computational resources required for accurate fraud detection compared to exhaustive rule-based analysis.
3Adaptability or versatility
If static fraud detection models are used, then model simplicity is maintained, but the system cannot adapt to the dynamic nature of cryptocurrency transactions
Solution Approach 1:
The patent implements dynamic fraud detection by continuously updating graph representations and behavioral features as new transactions occur. The system adapts to the dynamic nature of cryptocurrency transactions by recalculating risk scores based on evolving transaction patterns and network topology, enabling scalability without sacrificing detection accuracy.
Solution Approach 2:
The patent creates a universal fraud detection framework that combines graph-based features, behavioral-based features, and machine learning models into a single adaptable system. This multi-functional approach allows the same system to handle various types of fraudulent activities across different blockchain networks, reducing the need for multiple specialized models while maintaining high adaptability.
Data Source
AI summary
Methods, systems, and devices for iterative graph embedding of a blockchain network are described. A blockchain embedding service generates, using a graph representation of a blockchain network and a node embedding model, a first set of node embeddings for a first set of blockchain addresses using transaction data for the first set of blockchain addresses. The platform generates, using the graph representation, a second set of node embeddings for a second set of blockchain addresses associated with new transaction data. Generating the second set of node embeddings includes executing, for each node corresponding to blockchain address of the second set of blockchain addresses, a random walk across a set of nodes starting with the node using the transaction data for the set of nodes, inputting data resulting from the random walk into the node embedding model, and computing a risk score for each of the second set of blockchain addresses.