University of Manchester

The Automatic Bio Data Scientist: Using Artificial Intelligence to Automate Biomedical Modelling

Deadline: Open all year round
Self Funded

Project Description

Mathematical models of biological systems seek to specify the complex mechanistic relationships between many measurable quantities over space and time. However, as our capability to generate experimental data improves, our ability to conceive of appropriate mechanistic descriptions, encoded in mathematics, to describe these biological phenomena has become a bottleneck. There are now more questions and data than can be thoroughly addressed given the number of computational biologists available.

One solution is to adopt a pure machine learning to “learn biology” and unravel the complex dependencies automatically from data. Deep neural networks (NNs) have been at the forefront of such advances in recent years fuelled by advances in computational machinery that have enabled NNs to scale to the analysis of large datasets and provide versatility over standard models. However, despite some successes in a range of biomedical research applications, NNs are often derided for their “black box” discoveries, lack of interpretability and the need for unrealistic quantities of training data.

The “Automatic BioData Scientist” project is an ambitious attempt to develop interpretable machine-based intelligence tools that enables biomedical researchers to perform complex experiments using data. The student will investigate novel learning strategies that will enable the addition of more structure and constraints within deep NNs and transform these into physically realistic statistical models. The student will also collaborate with a range of Manchester-based biomedical scientists to identify complex, but common, computational analyses whose solution requires automation. This is an excellent opportunity for a machine learning researcher to make a substantial impact in the introduction of AI approaches to biomedical research design.

Funding Information

Applications are invited from self-funded students. This project has a Standard fee. Details of our different fee bands can be found on our website. For information on how to apply for this project, please visit the Faculty of Biology, Medicine and Health Doctoral Academy website.

As an equal opportunities institution we welcome applicants from all sections of the community regardless of gender, ethnicity, disability, sexual orientation and transgender status. All appointments are made on merit.

Eligibility Requirements

Candidates are expected to hold (or be about to obtain) a minimum upper second class honours degree (or equivalent) in a strongly mathematical subject (e.g. mathematics, statistics, engineering, physics, computer science). Candidates with experience in statistical modeling or machine learning and with an interest in biomedical applications are encouraged to apply.

Application Process

For information on how to apply for this project, please visit the Faculty of Biology, Medicine and Health Doctoral Academy website (bmh.manchester.ac.uk/study/research/apply/). Informal enquiries may be made directly to the primary supervisor. On the online application form select PhD Bioinformatics

For international students we also offer a unique 4 year PhD programme that gives you the opportunity to undertake an accredited Teaching Certificate whilst carrying out an independent research project across a range of biological, medical and health sciences. For more information please visit internationalphd.manchester.ac.uk

References

Martens, K., and Yau. C. (2020) Neural Decomposition: Functional ANOVA with Variational Autoencoders, AISTATS.

Martens, K., and Yau, C. (2020) Translation-invariant feature-level clustering with Variational Autoencoders, AISTATS.

Martens, K., Campbell, K. R., and Yau. C. (2019) Decomposing feature-level variation with Covariate Gaussian Process Latent Variable Models, International Conference on Machine Learning.

Campbell, K. C., and Yau, C. (2018) Uncovering genomic trajectories with heterogeneous genetic and environmental backgrounds from single-cell and bulk population data, Nature Comms.

Rukat, T., Holmes, C., and Yau. C. (2018) Probabilistic Tensor Factorisation, International Conference on Machine Learning.

Verified by MonsterInsights