A CTO is evaluating graph neural networks (GNNs) for a new use case. Which problem type represents the STRONGEST fit for GNN-based approaches compared to traditional deep learning?
-
A
Sequential time-series forecasting for stock prices
-
B
Fraud detection in transaction networks with complex relationship patterns
-
C
Image classification with high-resolution satellite imagery
-
D
Natural language summarization of long documents