This course is inspired by research papers that integrate Bayesian approaches with AI to analyze spatio-temporal data.

• Statistical Deep Learning for Spatial and Spatiotemporal Data (Annual Review of Statistics and Its Application, Christopher K. Wikle and Andrew Zammit-Mangion, 2023). 

• Comparison of Deep Neural Networks and Deep Hierarchical Models for Spatio-Temporal Data (Journal of Agricultural, Biological and Environmental Statistics, Christopher K. Wikle, 2017). 

• Scalable spatiotemporal prediction with Bayesian neural fields (Nature Communications, Saad et. al, 2024). 

• Variable selection of nonparametric spatial autoregressive models via deep learning (Spatial Stat, Xiaodi Zhang & Yunquan Song, 2024)

This course will be taught by an instructor with a background in statistics. We will begin with a basic introduction to neural networks and machine learning approaches, tailored for students without prior experience in these areas. Next, we will introduce Python—the most widely used programming language in AI—covering essential features relevant to our applications, though not in a comprehensive or exhaustive manner. We will then use the PyTorch library to perform computations on real examples. Students with a statistics background will learn how to implement PyTorch for parameter estimation, while those without a statistics background will gain exposure to statistical modeling techniques that differ from typical AI approaches.

• The coding exercises and examples will be based on the following playlist:

  • https://www.youtube.com/playlist?list=PLqnslRFeH2UrcDBWF5mfPGpqQDSta6VK4

• If you are not familiar with Python, please take some time to learn the basics from online resources. I will only provide a brief introduction in class. A good starting point is this Python tutorial playlist:

  • https://www.youtube.com/playlist?list=PLmvLOUyx8RK9DBZ1NgV4RoNDu4hrWqXQV

Motivation & Purpose to Teach This Course

Tentative Syllabus

Regression: https://machinelearningmastery.com/building-a-regression-model-in-pytorch/

https://machinelearningmastery.com/start-here/#pytorch

ResNet: https://medium.com/@chen-yu/building-a-customized-residual-cnn-with-pytorch-471810e894ed