Skills need an upgrade as digital techniques take hold

Share this on social media:

By Richard Kidd, Head of Chemistry Data at the Royal Society of Chemistry

Our understanding of the universe and scientific research are inexorably linked, of that there is no debate. As we improve our knowledge in one area, the other inevitably benefits.

It follows then that if we are to improve our understanding of the universe, we must strive to improve our capabilities. That means constantly reassessing whether the tools and techniques available to us through those advances can augment – or even reshape – research capability.

One area in which there has been significant recent progress is machine learning and artificial intelligence (AI). Just five or 10 years ago, the thought of incorporating these technologies into a research project may have been greeted with a shudder down the spine, straightforward cynicism or even complete dismissal of the notion.

The reluctance was understandable at the time, with limited computing power and the fledgling nature of AI producing projects that were, at best, subject to considerable trial and error. Racist chatbots, unsafe treatment recommendations from a supercomputer and incorrectly identifying congresspeople as wanted criminals served as testimony of work to be done and the dangers of underdeveloped approaches and software.

Today’s picture is entirely different. Machine learning and AI have developed to such an extent that the sight of a fully automated robot conducting experiments of its own accord is no longer science-fiction – it can in fact be seen in action at the University of Liverpool’s Cooper Group.

There are other well-known examples. From Lee Cronin’s ‘Chemputer’ at the University of Glasgow to Purdue University’s recent announcement on creating a library of chemical reactions to aid drug discovery, digital techniques are here to stay.

These developments exist at the nexus between the sheer computing power of computers and human ingenuity. Used properly, they can supercharge innovation.

Text and Data Mining is perhaps the perfect example of this. Using libraries such as those offered by the Royal Society of Chemistry, for example, research teams now have the opportunity to mine generations of human endeavour – even across disciplines – and find answers to questions we may have been asking for decades. With computers unmatched at finding patterns across wide ranges of data, we could even find ourselves asking entirely new questions and opening new avenues for research.

Realising the potential of this digital future is very exciting. There is however a bottleneck. The conventional research team is simply unequipped to take full advantage of these tools – with few of us able to programme a computer to carry out the myriad of tasks they could do to enhance our research capability.

With digital techniques advancing more every day and showing ever more worth to research teams, there is a growing urgency to address this shortcoming sooner rather than later. From an educational perspective, graduates across scientific disciplines will have to be familiar with the language of AI and machine learning to better plan research projects.

Programmers and technicians will become an increasingly important part of research teams, but it is also important they speak the language of the researcher. This will be essential to building teams with the diverse range of skills required in this new frontier of research.

There is no substitute for the ingenuity or imagination of human scientists, but the prospect of research accelerated by a new wave of smart technology is one that whets the appetite and poses that most tantalising question – what can we discover next?

More information on the Royal Society of Chemistry’s Text and Data Mining service can be found at: https://rsc.li/tdm

This article is brought to you by: 
Richard Kidd