1 code implementation • 5 Feb 2024 • David R. Burt, Yunyi Shen, Tamara Broderick
Unfortunately, classical approaches for validation fail to handle mismatch between locations available for validation and (test) locations where we want to make predictions.
1 code implementation • 20 Feb 2023 • Renato Berlinghieri, Brian L. Trippe, David R. Burt, Ryan Giordano, Kaushik Srinivasan, Tamay Özgökmen, Junfei Xia, Tamara Broderick
Given sparse observations of buoy velocities, oceanographers are interested in reconstructing ocean currents away from the buoys and identifying divergences in a current vector field.
no code implementations • 4 Nov 2022 • Vidhi Lalchand, Wessel P. Bruinsma, David R. Burt, Carl E. Rasmussen
In this work we propose an algorithm for sparse Gaussian process regression which leverages MCMC to sample from the hyperparameter posterior within the variational inducing point framework of Titsias (2009).
1 code implementation • 14 Oct 2022 • Alexander Terenin, David R. Burt, Artem Artemev, Seth Flaxman, Mark van der Wilk, Carl Edward Rasmussen, Hong Ge
For low-dimensional tasks such as geospatial modeling, we propose an automated method for computing inducing points satisfying these conditions.
no code implementations • 15 May 2022 • Andrew Y. K. Foong, Wessel P. Bruinsma, David R. Burt
The Chernoff bound is a well-known tool for obtaining a high probability bound on the expectation of a Bernoulli random variable in terms of its sample average.
1 code implementation • 23 Feb 2022 • Beau Coker, Wessel P. Bruinsma, David R. Burt, Weiwei Pan, Finale Doshi-Velez
Finally, we show that the optimal approximate posterior need not tend to the prior if the activation function is not odd, showing that our statements cannot be generalized arbitrarily.
no code implementations • NeurIPS Workshop ICBINB 2021 • David R. Burt, Artem Artemev, Mark van der Wilk
We suggest a method that adaptively selects the amount of computation to use when estimating the log marginal likelihood so that the bias of the objective function is guaranteed to be small.
1 code implementation • NeurIPS 2021 • Andrew Y. K. Foong, Wessel P. Bruinsma, David R. Burt, Richard E. Turner
Interestingly, this lower bound recovers the Chernoff test set bound if the posterior is equal to the prior.
no code implementations • 16 Feb 2021 • Artem Artemev, David R. Burt, Mark van der Wilk
We propose a lower bound on the log marginal likelihood of Gaussian process regression models that can be computed without matrix factorisation of the full kernel matrix.
2 code implementations • pproximateinference AABI Symposium 2021 • David R. Burt, Sebastian W. Ober, Adrià Garriga-Alonso, Mark van der Wilk
Then, we propose (featurized) Bayesian linear regression as a benchmark for `function-space' inference methods that directly measures approximation quality.
1 code implementation • 1 Aug 2020 • David R. Burt, Carl Edward Rasmussen, Mark van der Wilk
Gaussian processes are distributions over functions that are versatile and mathematically convenient priors in Bayesian modelling.
no code implementations • 23 Jun 2020 • David R. Burt, Carl Edward Rasmussen, Mark van der Wilk
We present a construction of features for any stationary prior kernel that allow for computation of an unbiased estimator to the ELBO using $T$ Monte Carlo samples in $\mathcal{O}(\tilde{N}T+M^2T)$ and in $\mathcal{O}(\tilde{N}T+MT)$ with an additional approximation.
no code implementations • 28 Jan 2020 • David Janz, David R. Burt, Javier González
We consider the problem of optimising functions in the reproducing kernel Hilbert space (RKHS) of a Mat\'ern kernel with smoothness parameter $\nu$ over the domain $[0, 1]^d$ under noisy bandit feedback.
2 code implementations • NeurIPS 2020 • Andrew Y. K. Foong, David R. Burt, Yingzhen Li, Richard E. Turner
While Bayesian neural networks (BNNs) hold the promise of being flexible, well-calibrated statistical models, inference often requires approximations whose consequences are poorly understood.
1 code implementation • 8 Mar 2019 • David R. Burt, Carl E. Rasmussen, Mark van der Wilk
Excellent variational approximations to Gaussian process posteriors have been developed which avoid the $\mathcal{O}\left(N^3\right)$ scaling with dataset size $N$.