Graph Analytic Systems
summary
Graph analytics gets applied to anti‑money‑laundering at Capital One by running large distributed computations (GraphX, GraphFrames, Apache Giraph, Flink) to do label propagation and reputation scoring, then loading results into a graph database (Neo4j/JanusGraph or AWS Neptune) for human queries. Bulk synchronous parallel (BSP) message‑passing and random‑walk based subgraph sampling feed neural embeddings so teams can run graph convolutional networks for community detection and “more‑like‑this” searches; the pattern: compute at scale, serve interactively.
AI-written summary · about
chapters (10)
- 0:00Introduction to talk
- 1:01Graph analytics in ML
- 5:43Use cases for graph analytics
- 11:41Graph processing engines
- 18:07Graph databases overview
- 27:43Graphical models in ML
- 35:48Query languages for graphs
- 40:27Graph neural networks
- 53:08Architectural summary
- 54:53Closing and resources
related talks
27mEnhancing Financial Analytics With Event-Driven Architectures and Graph Neural Networks
49mModern Graph Databases
26mAdvancing Financial Fraud Detection with Graph Dbs
43mAnalytics and AI Applied to Graph Databases - Introduction and Demo
23mGraph ML – The Next Level Of Machine Learning; Learn How To Use It And When Not To
45mApplications of Graph in Fraud Detection
26mLarge Scale Graph Processing & Machine Learning Algorithms
31mLet's do graph machine learning on Cloud