skip to content

Cambridge Service Alliance

At the forefront of service transformation in the digital era
 
December 2018 Paper

December 2018 Paper on 'Bayesian Neural Network Ensembles'

December 2018 Paper on 'Bayesian Neural Network Ensembles' by Tim Pearce, Mohamed Zaki and Andy Neely

Ensembles of neural networks (NN) have long been used to estimate predictive uncertainty; a small number of NNs are trained from different initialisations and sometimes on differing versions of the dataset. The variance of the ensemble’s predictions is interpreted as its epistemic uncertainty. The appeal of ensembling stems from being a collection of regular NNs - this makes them both scalable and easily implementable. NN ensembles have continued to achieve strong empirical results in recent years. The departure from Bayesian methodology is of concern since the Bayesian framework provides a principled, widely-accepted approach to handling uncertainty. This paper considered a method to produce Bayesian behaviour in neural networks (NN) ensembles by leveraging randomised MAP sampling. It departs only slightly from the usual handling of NNs, with parameters regularised around values drawn from a prior distribution. We showed that for NNs of sufficient width, each produces a sample from the posterior predictive distribution. Qualitative and benchmarking experiments were encouraging. Our ongoing work considers extending the presented theory to classification tasks as well as other architectures such as convolutional NNs.

[Paper]

Cambridge Service Alliance

Welcome to the Cambridge Service Alliance…

  • a unique global alliance between leading businesses and universities;
  • bringing together the world's leading firms and academics;
  • all of whom are devoted to delivering today the tools, education and insights needed for the complex service solutions of tomorrow.

LinkedIn

Follow us on Twitter

Latest news

Bouygues Construction joins the Cambridge Service Alliance

27 May 2021

We are delighted to announce that Bouygues Construction has joined the Cambridge Service Alliance, Institute for Manufacturing (IfM), University of Cambridge as a new industry member alongside our existing partners HCL Technologies , Caterpillar Inc. , Perkins Engines Company Limited and CEMEX . Press release

'Using AI to Track How Customers Feel — In Real Time' published in the Harvard Business Review

13 May 2021

The paper 'Using AI to Track How Customers Feel — In Real Time' by Mohamed Zaki, Janet R. McColl-Kennedy and Andy Neely has been published in Harvard Business Review.

Industry Day 2020 | What we learnt

6 December 2020

What is the future of services in the new pandemic-driven era? This was the topic explored at our 2020 Industry Day by senior executives from L'Oréal, Microsoft Research, Manchester United and start-up, Fairjungle.