Advanced Statistical Computing

Graduate credit

This course surveys essential tools and techniques for improving the speed, stability, and usability of research code. Topics include vectorization and tensor computations; automatic differentiation and just-in-time compilation; debugging, optimizing, and testing statistical procedures; essentials of GPU computing and CUDA; parallel and distributed computing; and recent advances in large-scale inference. Students are expected to contribute their own research problems, which will serve as the basis for a class project.