no code implementations • 8 Sep 2023 • Erik Daxberger, Floris Weers, BoWen Zhang, Tom Gunter, Ruoming Pang, Marcin Eichner, Michael Emmersberger, Yinfei Yang, Alexander Toshev, Xianzhi Du
We empirically show that our sparse Mobile Vision MoEs (V-MoEs) can achieve a better trade-off between performance and efficiency than the corresponding dense ViTs.
no code implementations • 17 Jun 2022 • Javier Antorán, David Janz, James Urquhart Allingham, Erik Daxberger, Riccardo Barbano, Eric Nalisnick, José Miguel Hernández-Lobato
The linearised Laplace method for estimating model uncertainty has received renewed attention in the Bayesian deep learning community.
no code implementations • pproximateinference AABI Symposium 2022 • Javier Antoran, James Urquhart Allingham, David Janz, Erik Daxberger, Eric Nalisnick, José Miguel Hernández-Lobato
We show that for neural networks (NN) with normalisation layers, i. e. batch norm, layer norm, or group norm, the Laplace model evidence does not approximate the volume of a posterior mode and is thus unsuitable for model selection.
no code implementations • 5 Nov 2021 • Runa Eschenhagen, Erik Daxberger, Philipp Hennig, Agustinus Kristiadi
Deep neural networks are prone to overconfident predictions on outliers.
3 code implementations • NeurIPS 2021 • Erik Daxberger, Agustinus Kristiadi, Alexander Immer, Runa Eschenhagen, Matthias Bauer, Philipp Hennig
Bayesian formulations of deep learning have been shown to have compelling theoretical properties and offer practical functional benefits, such as improved predictive uncertainty quantification and model selection.
no code implementations • NeurIPS 2021 • Erik Daxberger, Agustinus Kristiadi, Alexander Immer, Runa Eschenhagen, Matthias Bauer, Philipp Hennig
Bayesian formulations of deep learning have been shown to have compelling theoretical properties and offer practical functional benefits, such as improved predictive uncertainty quantification and model selection.
1 code implementation • 28 Oct 2020 • Erik Daxberger, Eric Nalisnick, James Urquhart Allingham, Javier Antorán, José Miguel Hernández-Lobato
In particular, we implement subnetwork linearized Laplace as a simple, scalable Bayesian deep learning method: We first obtain a MAP estimate of all weights and then infer a full-covariance Gaussian posterior over a subnetwork using the linearized Laplace approximation.
no code implementations • pproximateinference AABI Symposium 2021 • Erik Daxberger, Eric Nalisnick, James Allingham, Javier Antoran, José Miguel Hernández-Lobato
In particular, we develop a practical and scalable Bayesian deep learning method that first trains a point estimate, and then infers a full covariance Gaussian posterior approximation over a subnetwork.
1 code implementation • NeurIPS 2020 • Austin Tripp, Erik Daxberger, José Miguel Hernández-Lobato
We introduce an improved method for efficient black-box optimization, which performs the optimization in the low-dimensional, continuous latent manifold learned by a deep generative model.
Ranked #1 on Molecular Graph Generation on ZINC
no code implementations • 11 Dec 2019 • Erik Daxberger, José Miguel Hernández-Lobato
Despite their successes, deep neural networks may make unreliable predictions when faced with test data drawn from a distribution different to that of the training data, constituting a major problem for AI safety.
no code implementations • 2 Jul 2019 • Erik Daxberger, Anastasia Makarova, Matteo Turchetta, Andreas Krause
However, few methods exist for mixed-variable domains and none of them can handle discrete constraints that arise in many real-world applications.
no code implementations • 30 Jun 2018 • Yunpu Ma, Volker Tresp, Erik Daxberger
In this paper, we extend models for static knowledge graphs to temporal knowledge graphs.