Mapping Gaussian Process Priors to Bayesian Neural Networks

By Daniel Flam-Shepherd. 2017

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Danijar Hafner
2. Sergey Levine
3. Jiaxin Shi
4. Guodong Zhang
5. Dustin Tran
6. Yoshua Bengio
7. Yingzhen Li
8. Ruslan Salakhutdinov

Citations in the corpus

• Noise Contrastive Priors for Functional Uncertainty (2018, background): “More recently, Depeweg et al. (2018) introduce networks with latent variables in order to disentangle forms of uncertainty, and Flam-Shepherd et al. (2017) propose general-purpose weight priors based on approximating Gaussian processes.”
• Variational Implicit Processes (2018, methodology): “A very recent work [26] also minimizes the sleep phase KL divergence (3) but wrt. the BNN prior, and their goal is to regularize BNN priors and implant some smoothness properties of GPs to BNNs.”
• Functional Variational Bayesian Neural Networks (2019, methodology): “Flam-Shepherd et al. (2017) trained a BNN prior to mimic a GP prior, but they still required variational inference in weight space.”