Coventry University Featured PhD Programmes
Sheffield Hallam University Featured PhD Programmes
University College London Featured PhD Programmes

Deep generative modelling from the classic probabilistic perspective

This project is no longer listed on FindAPhD.com and may not be available.

Click here to search FindAPhD.com for PhD studentship opportunities
  • Full or part time
    Dr O Isupova
  • Application Deadline
    No more applications being accepted
  • Competition Funded PhD Project (European/UK Students Only)
    Competition Funded PhD Project (European/UK Students Only)

Project Description

The University of Bath is inviting applications for the following PhD project in the Department of Computer Science under the supervision of Dr Olga Isupova https://researchportal.bath.ac.uk/en/persons/olga-isupova.

Project overview:

In classic probabilistic modelling, generative models are considered in contrast to discriminative models. Generative models describe how data is generated and then categorise the data, rather than categorising the data directly, as in discriminative models.

In deep learning, generative models are considered mostly as a tool to generate realistic objects, such as realistic looking photos of high resolution created by Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs).

This project will research modern generative models from the classic probabilistic perspective. The idea is to challenge the choice of reconstruction loss as the main objective for generative models. If we want to build meaningful and useful data representation, we do not necessarily care that this representation can be used to perfectly reconstruct background pixels in an image. Generative models in traditional probabilistic modelling use objectives such as likelihood or mutual information maximisation. This project will investigate the potential of these and related objectives in deep generative models.

The project will have opportunities to collaborate with Unicef, Save The Elephants, department of Zoology and School of Geography and Environment at University of Oxford. The developed methods can be tested on data provided by the collaborators.

Candidate requirements:

Candidates are expected to have or be near completion of an MSc or MEng (or equivalent) in Computer Science, Mathematics, Statistics or related areas. A strong mathematical background and programming experience are desirable.

Enquiries and applications:

Informal enquiries are welcomed and should be directed to Dr Olga Isupova, [Email Address Removed].

Formal applications should be made via the University of Bath’s online application form for a PhD in Computer Science:
https://samis.bath.ac.uk/urd/sits.urd/run/siw_ipp_lgn.login?process=siw_ipp_app&code1=RDUCM-FP01&code2=0014

More information about applying for a PhD at Bath may be found here:
http://www.bath.ac.uk/guides/how-to-apply-for-doctoral-study/

Anticipated start date: 28 September 2020.

Funding Notes

Research Council funding is available for an excellent UK or EU student who has been ordinarily resident in the UK since September 2017. For more information on eligibility: https://www.epsrc.ac.uk/skills/students/help/eligibility/.

Funding will cover UK/EU tuition fees, maintenance at the UKRI doctoral stipend rate (£15,285 per annum tax-free in 2020/21, increasing annually in line with the GDP inflator) and a training support grant (£1,000 per annum) for a period of up to 3.5 years.

We also welcome all-year-round applications from self-funded candidates and candidates who can source their own funding.

References

Oord, A.V.D., Li, Y. and Vinyals, O., 2018. Representation learning with contrastive predictive coding.

How good is research at University of Bath in Computer Science and Informatics?

FTE Category A staff submitted: 24.00

Research output data provided by the Research Excellence Framework (REF)

Click here to see the results for all UK universities


FindAPhD. Copyright 2005-2020
All rights reserved.