AI for Finance 2022

馃嚞馃嚙 Blockchain Analysis with Graph Neural Networks

Sep 20, 2022 | 10:26 AM CEST - 10:51 AM CEST

Stage 2


How to detect fraudulent Ethereum blockchain transactions and accounts using a TigerGraph-powered Graph Neural Network? Graph is the most natural data model. Graph is how we think. Overall, Graph is now becoming a key component in AI/ML projects. This session will discuss how: 路 Native graph algorithms provide unsupervised learning solutions and anomaly detection capabilities 路 Graph features enrich training data with the contextual information 路 Graph data models are used for large-scale similarity search 路 Graph embeddings transform graph into easier-to-use vectors for traditional ML 路 Graph data models are used for explanatory AI 路 Graph visualizations are critical components of advanced investigations UI

Presented by

Sponsored by