Variational Inference for BNNs (e.g., Bayes by Backprop)
Was this section helpful?
Weight Uncertainty in Neural Networks, Charles Blundell, Julien Cornebise, Koray Kavukcuoglu, Daan Wierstra, 2015Proceedings of the 32nd International Conference on Machine Learning (ICML), Vol. 37DOI: 10.48550/arXiv.1505.05424 - Original paper describing the Bayes by Backprop algorithm for Bayesian Neural Networks, a cornerstone method for VI in deep learning.
Auto-Encoding Variational Bayes, Diederik P Kingma, Max Welling, 2014International Conference on Learning Representations (ICLR)DOI: 10.48550/arXiv.1312.6114 - Introduces the reparameterization trick, a method central to gradient-based optimization in variational inference, particularly for models like VAEs and BNNs.
Probabilistic Machine Learning: Advanced Topics, Kevin Patrick Murphy, 2023 (MIT Press) - An advanced textbook containing detailed chapters on Variational Inference and its use in Bayesian Neural Networks.