Academic Interview - Sara Wade
Taraneh latifi Seresht and Manjari Agrawal have worked together to produce this article as part of our series of Academic Interviews; featuring Sara Wade!
A researcher and lecturer in the fields of statistics and machine learning, Dr. Sara Wade is one of the leading female mathematicians at the University of Edinburgh. She grew up in Maryland, USA, with a budding interest in Mathematics, bolstered by the fact that her parents were actuaries.
The love of numbers was always present and, during her time as an undergraduate at the University of Maryland, she found her interests converging towards the more applied areas of the field. Furthering this newfound inclination, she took more courses in statistics. After this Sara did an internship at the US Census Bureau where she was introduced to Bayesian statistics, which is one of the fields she specialises in as a researcher.
Sara moved to Italy after meeting her husband, who hails from there, during her time as an undergraduate. She completed her PhD at the University of Milan, abode of the famous Bayesian statistician Bruno De Finetti. A drastic change, after staying close to home for the first part of her education, but a welcome one she thinks since she was able to collaborate with a small cohort of likeminded PhD students.
This led to Sara taking up a postdoc position at the University of Cambridge, whose machine learning group furthered her interests. From there she went on to become a lecturer at the University of Warwick, a good choice she felt since it had a separate department for statistics. This culminated in her current role as a lecturer at the University of Edinburgh. It was an option which appealed to her with its strong informatics and growing statistics departments which housed many Bayesian statisticians.
As a lecturer, she teaches a course called Machine Learning and Python which is offered to students doing their masters, or who are in their final year as an undergraduate. She designed this when she joined the University two years ago, and it has been on offer since. Talking about the course she says that 'there's a lot of interest in machine learning now', which seems to be the case since it has almost 200 students enrolled. In the beginning the approach she took with the class was based on the assumption that the students would have a higher level of background knowledge. Though she goes on to say that 'you learn and adapt along the way'; she did this by adjusting the course to suit students on different levels.
Sara talks about her current research project on scalar-on-regression models, which she's working on with a PhD student; for which she received the Royal Society of Edinburgh grant. The research is focused on diagnosing dementia and predicting the decline of cognitive skills over a couple of years due to the disease. At the moment a definite diagnosis can be only formed at autopsy. There are two extreme methods of evaluating the onset of the ailment. The first processes the thickness and volume of regions of the cortex to form the model; for the second we have complex learning methods which take in the whole image and slice it into patches and form a prediction. The latter's improved accuracy isn't perfect as 'it's very hard to understand what goes on behind the scenes'. At the moment they are trying to look in between the two extremes.
Reflecting on herarea of research, Sara believes it's a welcoming space for different people who have different strengths. If someone is coming in from a statistics background it's worth it 'as long as you can read the machine learning literature and incorporate that into what you're doing'. At the same time, coming from a machine learning background, with tools in statistics, can help you assess models or the algorithms formally. 'Even if you have a strength in statistics and machine learning, there's scope for you to combine some of the advances from both fields'.
Finally, exchanging International Women's day wishes, she talks about being a woman in Mathematics. Pursuing her interest in computer science as an undergraduate, she felt intimidated and backed away from it since she was the only woman in the course. Even at Cambridge when she decided to go for machine learning, she was the only female postdoc and the most senior one in the group. Sara's in the Women in Machine Learning group and being in a minority she thinks that it's helpful to find a network of other women in the same field. 'There's a network out there, so, even if you feel alone reach out and try to find it, just stick with it and push through it if it's something that you like, and you love to do'.
Ending on an enthusiastic note she observes the equal ratio in the faculty in the statistics department at the University which she believes is encouraging not just for her, but for her two young daughters as well!