Applications of machine learning to research by Calum Hand

Posted in: Doctoral Exchange, Research skills

A student blog as part of the Doctoral Exchange series organised by the Doctoral College.

Machine learning (ML) is an exciting area of mathematics that allows computers to learn trends present in large volumes of data and use them to: allow cars to drive autonomously, provide better recommendations for services like YouTube and Spotify, and dozens of other applications we use in day-to-day life.

As a PhD student however, ML can also be a powerful research tool, allowing for prior experimental data to form the bedrock of a predictive model which can rapidly accelerate the rate of research.

How to get started using Machine Learning

Unfortunately, it is often difficult to know where to start with ML, so I was delighted to lead a roundtable discussion with the Doctoral College focussing on the considerations interested researchers should make when utilising ML.

The format of the roundtable event was structured as a half-and-half blend of me presenting, and group discussion of several core questions that are vital considerations for any machine learning project:

  1. What problem are you specifically attempting to solve?
  2. Do you have data available to train an ML model based on this problem?
  3. Is your data biased or are there significant underlying assumptions?
  4. Do you know what particular bits of data you should use to describe your problem when training the ML model?
  5. Are you able to suitably measure the predictive capability of your ML model?

The issues faced by researchers at Bath using Machine Learning

Opening up the questions to the audience gave a fascinating insight into the research going on at Bath, and the various ways people were attempting to solve the above problems in fields ranging from Project Management to Biochemistry. It also allowed us to take stock and appreciate that machine learning is not quite the silver bullet it is often touted as by the research community, but is instead an application of the scientific method and as such equally deserving of critique and rigour.

If anyone is interested in learning more about applying machine learning to their research please reach out to me at so I can add you to the “Applied Machine Learning” interest group I have started at Bath.

About Doctoral Exchange

This blog was written as part of the Doctoral Exchange series, a round-table discussion series for doctoral researchers to share experiences and ideas in a peer-to-peer environment. All topics are student-led. The programme can be found online on the Doctoral Exchange webpage. If you are interested in facilitating a session then please email

Posted in: Doctoral Exchange, Research skills

Get in touch if you're interested in learning more about applying machine learning to your research


  • (we won't publish this)

Write a response