Episode
How Data Scientists Use Graph Neural Networks for Fraud Detection
- Podcast
- The Data Science Podcast with Fexingo: Analytics, Machine Learning, and Data-Driven Conversations
- Published
- Jul 15, 2026
- Duration seconds
- 580
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/the-data-science-podcast-with-fexingo-analytics-machine-learning-and-data-driven-conversations-7871831/episodes/how-data-scientists-use-graph-neural-networks-for-fraud-detection/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/the-data-science-podcast-with-fexingo-analytics-machine-learning-and-data-driven-conversations-7871831/how-data-scientists-use-graph-neural-networks-for-fraud-detection.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Episode 111 dives into graph neural networks (GNNs) for detecting fraud in financial transactions. Lucas and Luna explore how GNNs model relational patterns between accounts, merchants, and devices — catching fraud rings that traditional models miss. They walk through a real case from 2025 where a European bank used GNNs to reduce false positives by 30 percent while catching 22 percent more synthetic identity fraud. The conversation covers inductive vs. transductive learning, node classification, and the challenge of evolving graph structures. Perfect for data scientists looking to apply GNNs beyond social networks. #GraphNeuralNetworks #FraudDetection #MachineLearning #DataScience #FinancialFraud #SyntheticIdentity #NodeClassification #InductiveLearning #TransactionGraphs #GraphML #Python #PyTorchGeometric #NetworkAnalysis #Technology #FexingoBusiness #BusinessPodcast #DataSciencePodcast #MLinProduction Keep every episode free: buymeacoffee.com/fexingo