Adaptive delayed-acceptance pseudo-marginal random walk Metropolis

Duration: 30 mins 24 secs

Share this media item:

Embed this media item:

<iframe width="_width_" height="_height_" src="https://sms.cam.ac.uk/media/1706234/embed" frameborder="0" scrolling="no" allowfullscreen></iframe>

Choose size:

About this item

Available Formats

About this item

Description:	Sherlock, C (Lancaster University) Wednesday 23 April 2014, 14:55-15:30


Created:	2014-04-28 17:05
Collection:	Advanced Monte Carlo Methods for Complex Inference Problems
Publisher:	Isaac Newton Institute
Copyright:	Sherlock, C
Language:	eng (English)


Abstract:	Co-authors: Andrew Golightly (Newcastle University) and Daniel Henderson (Newcastle University) Delayed-acceptance (DA) pseudo-marginal Metropolis-Hastings (MH) algorithms can be applied when it is computationally expensive to calculate an unbiased estimate of the true posterior, but a computationally cheap approximation is available. A first accept-reject stage is applied, with the cheap approximation substituted for the true posterior in the MH acceptance ratio. Only for those proposals which pass through the first stage is the computationally expensive true posterior (or unbiased estimate thereof) evaluated, with a second accept-reject stage ensuring that detailed balance is satisfied with respect to the intended posterior. A weighted average of all previous unbiased estimates of the true posterior provides a generic approximation to the true posterior. If only the k-nearest neighbours are used in the average then evaluation of the approximate posterior can be made computationally cheap provided that the points at which the posterior has been estimated unbiasedly are stored in a multi-dimensional binary tree, similar to a KD-tree.

Abstract:

Co-authors: Andrew Golightly (Newcastle University) and Daniel Henderson (Newcastle University)
Delayed-acceptance (DA) pseudo-marginal Metropolis-Hastings (MH) algorithms can be applied when it is computationally expensive to calculate an unbiased estimate of the true posterior, but a computationally cheap approximation is available. A first accept-reject stage is applied, with the cheap approximation substituted for the true posterior in the MH acceptance ratio. Only for those proposals which pass through the first stage is the computationally expensive true posterior (or unbiased estimate thereof) evaluated, with a second accept-reject stage ensuring that detailed balance is satisfied with respect to the intended posterior. A weighted average of all previous unbiased estimates of the true posterior provides a generic approximation to the true posterior. If only the k-nearest neighbours are used in the average then evaluation of the approximate posterior can be made computationally cheap provided that the points at which the posterior has been estimated unbiasedly are stored in a multi-dimensional binary tree, similar to a KD-tree.

Available Formats

Format	Quality	Bitrate	Size
MPEG-4 Video	640x360	1.93 Mbits/sec	441.05 MB	View	Download
WebM	640x360	408.84 kbits/sec	91.08 MB	View	Download
iPod Video	480x270	522.08 kbits/sec	116.25 MB	View	Download
MP3	44100 Hz	249.76 kbits/sec	55.67 MB	Listen	Download
Auto *	(Allows browser to choose a format it supports)