Get our free extension to see links to code for papers anywhere online!

Add to Chrome

Add to Firefox

Picture for José Miguel Hernández-Lobato

José Miguel Hernández-Lobato

Adapting the Linearised Laplace Model Evidence for Modern Deep Learning

Jun 17, 2022
Javier Antorán, David Janz, James Urquhart Allingham, Erik Daxberger, Riccardo Barbano, Eric Nalisnick, José Miguel Hernández-Lobato

Figure 1 for Adapting the Linearised Laplace Model Evidence for Modern Deep Learning

Figure 2 for Adapting the Linearised Laplace Model Evidence for Modern Deep Learning

Figure 3 for Adapting the Linearised Laplace Model Evidence for Modern Deep Learning

Figure 4 for Adapting the Linearised Laplace Model Evidence for Modern Deep Learning

* Paper appearing at ICML 2022

Access Paper or Ask Questions

Meta-learning Feature Representations for Adaptive Gaussian Processes via Implicit Differentiation

May 05, 2022
Wenlin Chen, Austin Tripp, José Miguel Hernández-Lobato

Figure 1 for Meta-learning Feature Representations for Adaptive Gaussian Processes via Implicit Differentiation

Figure 2 for Meta-learning Feature Representations for Adaptive Gaussian Processes via Implicit Differentiation

Figure 3 for Meta-learning Feature Representations for Adaptive Gaussian Processes via Implicit Differentiation

Figure 4 for Meta-learning Feature Representations for Adaptive Gaussian Processes via Implicit Differentiation

* 17 pages, 6 figures, 3 tables, 1 algorithm

Access Paper or Ask Questions

Addressing Bias in Active Learning with Depth Uncertainty Networks... or Not

Dec 13, 2021
Chelsea Murray, James U. Allingham, Javier Antorán, José Miguel Hernández-Lobato

Figure 1 for Addressing Bias in Active Learning with Depth Uncertainty Networks... or Not

Figure 2 for Addressing Bias in Active Learning with Depth Uncertainty Networks... or Not

Figure 3 for Addressing Bias in Active Learning with Depth Uncertainty Networks... or Not

Figure 4 for Addressing Bias in Active Learning with Depth Uncertainty Networks... or Not

* arXiv admin note: substantial text overlap with arXiv:2112.06796

Access Paper or Ask Questions