Explore scalable graph machine learning techniques in this one-hour conference talk by Viktor Prasanna, recorded for SPCL_Bcast #44 on November 9, 2023. Delve into the computational challenges of Graph Neural Networks (GNNs) and their applications. Learn about innovative approaches developed at USC's Data Science Lab and FPGA/Parallel Computing Lab, including GraphSAINT for efficient graph embedding and graph processing over partitions (GPOP) for handling large-scale graphs. Discover how FPGA implementations achieve significant speedups compared to CPU and GPU counterparts. Examine specific accelerators for GraphSAGE and GraphSAINT models, and gain insights into opportunities and challenges in leveraging heterogeneous architectures for GNN acceleration. For more talks in this series, visit the SPCL Bcast website.