Case study: Alex Robertson

Alex joined the Engineering and Physical Sciences Research Council-supported CDT for Data Science at the University of Edinburgh in 2016, after completing an undergraduate degree in linguistics at the University of Cambridge and work as a research assistant in psycholinguistics at the University of Edinburgh. His research focuses on the emergent linguistic properties of emoji on social media, looking at how people both use and perceive emoji, how this changes based on context and social factors, and how this all changes over time.

“I have access to excellent computational resources at the CDT, which is important since I work with enormous amounts of social media data. Furthermore, my supervisors are highly experienced researchers who are supportive of my academic endeavours. There has also been a lot of support available for travelling to conferences in order to present my work. Travel funding is another benefit. There are also many events organised, which means a very cohesive social and professional experience here - though I work on my own project, I feel part of a wider group of researchers and I can turn to them for support when needed.

“My research is important because although emoji are increasingly common in online communication, there is very little research into their linguistic properties. Understanding how language-like emoji are will help us to understand how properties of language emerge and how these vary based on a variety of social factors such as age, gender and ethnicity.

Language-based AI generally ignore emoji. My work has already shown that emoji carry significant information about the author's identity. So there is scope for using emoji as features for AI models to focus on. For example, I have already trained classifiers that can predict author gender or ethnicity fairly well, using only the emoji that person has used on social media.

“I'm fortunate to be able to research something I am passionate about but have also gained a lot of new skills. Not only technical but more "soft" skills. These will stand me in good stead for whatever career path I take in the future. And the quality of my training will mean that I have many paths to choose from, so I feel very optimistic about my post-PhD opportunities.

“I expect it will streamline the progression for future PhD applicants, if MSc and PhD are combined as in my CDT. That is good for applicants. More broadly, giving a wide range of people the freedom to work on a wide variety of projects related to AI can really only be a good thing and encourage innovation.”