£6,000 PhD Scholarship | APPLY NOW £6,000 PhD Scholarship | APPLY NOW

Neural Network Descramblers: A New Approach to Model Interpretation


   Department of Chemistry

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

Click here to search FindAPhD.com for PhD studentship opportunities
  Dr Matthew Grayson  No more applications being accepted  Competition Funded PhD Project (UK Students Only)

About the Project

Collaborative Training Partnership in Innovation:

The University of Bath and Corsham Science Ltd have established a unique Collaborative Training Partnership in Innovation and are inviting applications for the following 4-year PhD project, commencing in October 2022. This partnership will deliver a bespoke training and support package to students embedding Innovation and Business into doctoral student cohorts, developing future research leaders in Industry, Academia and at the interface.

In addition to the significantly enhanced funding package, students will also be provided training courses in business innovation, annual industry-led conferences, project management training, access to industry R&D facilities and staff, and an industrial training placement (3-6 months).

Supervisory Team:

Overview of the Research:

We have an exciting opportunity to work on the development of new methods for neural network interpretation. You will join the University of Bath to develop the use of neural network descramblers to create more transparent models of chemical properties relative to traditional machine learning methods.

As a PhD student, you will:

  • Engage in training in coding and machine learning.
  • Gain experience in building neural networks and developing methods for their interpretation.
  • Perform both individual and collaborative research projects.
  • Write up research results for publication in scientific journals.
  • Disseminate your work through presentations at national and international conferences.

Project Details

Black-box neural networks (NNs) offer limited insights into how they work and thus their widespread adoption in the biological and physical sciences remains challenging in the absence of such transparency. A recent work (DOI: 10.1073/pnas.2016917118) has proposed a method of "descrambling" NNs to extract the underlying mathematical functions that these models have learnt. This method offers a promising means to interpret the inner workings of NNs and increase their transparency, thus overcoming some of the current limitations of deep NNs. This project proposes the use of NN descramblers to create systems for predicting chemical properties in a reliable and explainable manner.

This will be achieved by training NNs on chemical datasets and applying the descrambling method on the networks to elucidate the underlying functions that were learned to predict the chemical properties. Since the extracted functions are less complex than the function of a NN, they are more easily inspected and understood. Therefore, an expert will be able to examine these functions and validate whether they are correctly describing the physics of the chemical problem. Following their validation, these learned functions may then be used to make predictions of the chemical properties that they describe, instead of the NN. These functions will be much more interpretable than a black-box NN and should have the same predictive performance.

Descrambling NNs holds the promise of more transparent and robust programs for chemical property prediction, and this project aims to define and expand the range of applicability of this method.

Project keywords: Machine learning, Neural networks, Data science, Computational chemistry, Data analysis, Interpretability.

Candidate Requirements:

We are looking for a highly motivated individual to join our team. You should hold, or expect to receive, a First Class or good Upper Second Class Honours degree (or the equivalent). A master’s level qualification would also be advantageous. Experience with coding (any language) and machine learning is desirable but not essential.

Non-UK applicants must meet our English language entry requirement.

Enquiries and Applications:

Informal enquiries are welcomed and should be directed to Dr Matthew Grayson on email address [Email Address Removed].

Formal applications should be made via the University of Bath’s online application form for a PhD in Chemistry.

More information about applying for a PhD at Bath may be found on our website

NOTE: Applications will be accepted up until the deadline (19 June), although shortlisting and arrangements for interviews will begin immediately upon receipt of application, so candidates are encouraged to complete the application process as soon as possible.

Funding Eligibility:

To be eligible for funding, you must qualify as a Home student. The eligibility criteria for Home fee status are detailed and too complex to be summarised here in full; however, as a general guide, the following applicants will normally qualify subject to meeting residency requirements: UK nationals (living in the UK or EEA/Switzerland), Irish nationals (living in the UK or EEA/Switzerland), those with Indefinite Leave to Remain and EU nationals with pre-settled or settled status in the UK under the EU Settlement Scheme). This is not intended to be an exhaustive list. Additional information may be found on our fee status guidance webpage, on the GOV.UK website and on the UKCISA website

International candidates will only be considered if they can provide evidence of their ability to fund the difference between the Home and Overseas tuition fee rates. For the 2022/23 academic year, the difference payable would be £20,800 and this figure could increase by up to 5% each year for every further year of study.

Equality, Diversity and Inclusion:

We value a diverse research environment and aim to be an inclusive university, where difference is celebrated and respected. We welcome and encourage applications from under-represented groups.

If you have circumstances that you feel we should be aware of that have affected your educational attainment, then please feel free to tell us about it in your application form. The best way to do this is a short paragraph at the end of your personal statement.


Funding Notes

The successful candidate will receive a 4-year studentship funded jointly by the University of Bath and Corsham Science Ltd (CSL). The studentship will cover Home tuition fees and will provide an annual stipend (UKRI rate of £16,062 in 2022/23) as well as a budget for consumables/research expenses, training, travel and conference attendance. Furthermore, the successful candidate could be eligible to receive an annual payment of at least £1,400 directly from CSL, who will also cover costs associated with an industrial placement (likely 3-6 months) in the company. Eligibility criteria apply – see Funding Eligibility section above.

How good is research at University of Bath in Chemistry?


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

Click here to see the results for all UK universities
Search Suggestions
Search suggestions

Based on your current searches we recommend the following search filters.

PhD saved successfully
View saved PhDs