Lightning Talk: Bayesian Neural Networks With Variational Inference in PyTorch - Lars Heyen

This talk covers the implementation of Bayesian Neural Networks (BNNs) using Variational Inference in PyTorch, emphasizing the significance of uncertainty quantification in critical applications of neural networks. The speaker discusses the challenges of traditional BNN implementations and introduces torch_blue, a lightweight open source library designed to facilitate easy and flexible research on BNNs. Attendees will gain insights into making BNNs more accessible while maintaining their inherent capability to quantify uncertainty.