1 code implementation • 1 Nov 2023 • Tennessee Hickling, Dennis Prangle
We propose a transformation capable of altering the tail properties of a distribution, motivated by extreme value theory, which can be used as a layer in a normalizing flow to approximate multivariate heavy tailed distributions.
1 code implementation • 22 Oct 2020 • Matthew A. Fisher, Tui Nolan, Matthew M. Graham, Dennis Prangle, Chris J. Oates
Measure transport underpins several recent algorithms for posterior approximation in the Bayesian context, wherein a transport map is sought to minimise the Kullback--Leibler divergence (KLD) from the posterior to the approximation.
1 code implementation • 12 Jun 2020 • Rob Geada, Dennis Prangle, Andrew Stephen McGough
One-shot Neural Architecture Search (NAS) aims to minimize the computational expense of discovering state-of-the-art models.
Ranked #39 on Neural Architecture Search on CIFAR-10
1 code implementation • 8 Oct 2019 • Dennis Prangle, Cecilia Viscardi
The training data is "distilled" by using it to train an updated normalizing flow.
1 code implementation • 2 Oct 2019 • Tom Ryder, Dennis Prangle, Andrew Golightly, Isaac Matthews
Variational inference has had great success in scaling approximate Bayesian inference to big data by exploiting mini-batch training.
no code implementations • 21 Jun 2019 • Wil O. C. Ward, Tom Ryder, Dennis Prangle, Mauricio A. Álvarez
Latent force models are systems whereby there is a mechanistic model describing the dynamics of the system state, with some unknown forcing term that is approximated with a Gaussian process.
1 code implementation • 5 Jun 2019 • Christopher Drovandi, Richard G. Everitt, Andrew Golightly, Dennis Prangle
Particle Markov chain Monte Carlo (pMCMC) is now a popular method for performing Bayesian statistical inference on challenging state space models (SSMs) with unknown static parameters.
Computation Methodology
2 code implementations • 11 Apr 2019 • Sophie Harbisher, Colin S Gillespie, Dennis Prangle
Most computational approaches to Bayesian experimental design require making posterior calculations repeatedly for a large number of potential designs and/or simulated datasets.
Computation
no code implementations • 20 Nov 2018 • Tom Ryder, Andrew Golighty, A. Stephen McGough, Dennis Prangle
State-space models (SSMs) provide a flexible framework for modelling time-series data.
2 code implementations • ICML 2018 • Thomas Ryder, Andrew Golightly, A. Stephen McGough, Dennis Prangle
Parameter inference for stochastic differential equations is challenging due to the presence of a latent diffusion process.
no code implementations • 12 Oct 2017 • Richard G. Everitt, Dennis Prangle, Philip Maybank, Mark Bell
Bayesian inference for models that have an intractable partition function is known as a doubly intractable problem, where standard Monte Carlo methods are not applicable.
no code implementations • 27 Apr 2016 • Dennis Prangle, Richard G. Everitt
We show that the auxiliary variable method (M{\o}ller et al., 2006; Murray et al., 2006) for inference of Markov random fields can be viewed as an approximate Bayesian computation method for likelihood estimation.
1 code implementation • 3 Jul 2015 • Dennis Prangle
Approximate Bayesian computation performs approximate inference for models where likelihood computations are expensive or impossible.
Computation
no code implementations • 21 Jan 2015 • Dennis Prangle
ABC algorithms involve a large number of simulations from the model of interest, which can be very computationally costly.