Professor Andrew Caines is a Research Professor in the Natural Language and Information Processing Group at the Department of Computer Science and Technology, and a member of the Automated Language Teaching and Assessment (ALTA) Institute since its foundation in 2013.
His research develops automated methods for teaching and assessing language, mainly for people learning English, with the aim of making assessment fairer and good teaching more widely available. He completed his PhD in Cambridge in 2010 and joined the University staff in October 2013.
This article was written by Shrankhla Pandey, a PhD student at the Department of Psychiatry and the Department of Computer Science and Technology.
A network before it was a fund
Andrew's first contact with Cambridge Language Sciences (CLS) was not a grant. It was a talk. He attended his first CLS event before he was even on the University staff, and what struck him was simple but profound.
"No other group regularly organises multi-disciplinary talks. This is refreshing, thought-provoking, and a way to guard against tunnel vision."
For Andrew, this is the IRC’s fundamental value, alongside something more practical: organised events at which colleagues from different faculties can regroup and make connections that might not otherwise happen.
The first tangible benefit came from a phone call, not a cheque
Before Andrew had received any funding from CLS, the network had already inspired the direction of his work. The CLS coordinator at that time, Jane Walsh, put his PI, Professor Paula Buttery, in touch with the Cambridge-Africa Programme, which was looking for mentors. That introduction led to a collaboration with Dr Fridah Katushemererwe of Makerere University in Uganda, two visits by Fridah to Cambridge, two visits by Andrew to Uganda, and a two-day workshop on computational linguistics. One product of that work stands out: a working spell-checker for the Runyakitara languages of western Uganda, languages which are mainly spoken rather than written and had no standardised spelling. Andrew notes that: "None of this would have happened if it weren't for CLS and Jane putting people in contact with each other."
Same fund, ten years, different roles
Andrew applied for his first Incubator Fund award in 2016, the year the Fund was launched. Since then, he has been awarded funding twelve times. In 2016 he was the early-career researcher with an idea and a collaborator in another department. In more recent years the awards are led by PhD students and research associates in his group, with Andrew and Paula Buttery as the senior names. He has seen the fund from different sides: as the junior researcher, it gave him an early chance; as the collaborator, it brought him into new work; and now, as the supervisor, he uses it to give the same opportunity to budding researchers.
Counting the collaborators on those awards, Andrew has worked across six University departments (Computer Science and Technology, Theoretical and Applied Linguistics, Spanish and Portuguese, Education, Engineering and Psychology) and one external partner (the AI research lab EleutherAI). Would those collaborations, and by extension the research, exist without CLS and its Incubator Fund? "It would be very tricky, if not impossible, for me to have achieved this without the Incubator awards."
Three things such small grants are good for
The Incubator Fund awards are modest: up to £2,500 when the Fund launched in 2016, now up to £5,000 per project. Looking across his ten years of awards, the money does three different jobs:
For undergraduates, it funds a summer. Most of Andrew's awards paid for a summer research internship. A student gets eight to ten weeks of real research alongside their degree, a report with their name on it, and sometimes a change of direction. One intern from 2018 went on to present his own work at the CLS Annual Symposium the following year and helped plan a master's in computational linguistics.
For early-career researchers, it breaks down potential barriers. Departments and faculties are not built for someone in Computer Science to work with someone in Spanish and Portuguese, or Education, or Engineering. A small joint grant is the excuse both sides need, and the shared task that turns an introduction into a working relationship.
For senior researchers, it is a low-risk testing ground. An idea that is too early or too unusual for a large grant can be tried for a few thousand pounds, and either it produces the evidence to go further or it does not, and little is lost.
Several of Andrew's awards became papers at leading conferences; others became datasets, tools or working groups; some simply answered a question. All of those are useful outcomes for the price.
More than money
The fund is easy to describe in terms of grants and their outputs, but the network around it matters just as much. CLS gave Andrew a platform to initiate the Cambridge Language Sciences Symposium for Early-Career Researchers in 2018, modelled on the Annual Symposium but aimed at postgraduate students and postdocs as a friendly place to present. Andrew led the event as Programme Organiser in its early years, with over 150 people attending across the events.
The network's sub-groups on Education in Challenging Contexts and on Language, Brains and Machines brought together people with intersecting interests who would not otherwise have met. And for an interdisciplinary institute like ALTA, CLS provides a way to explore the edges of its remit and bring new findings back into the mainstream of its work.