The final stage of the 2nd edition of the Data Science Game was held last week at Capgemini’s Les Fontaines campus near Paris. The gathering saw 80 students from around the world compete against each other - and the UK team from Cambridge came second, behind the Moscow Institute of Physics and Technology.
Cantab, the University of Cambridge team, comprised Max Chamberlin (from Mill Hill, London), Kostas Tsakalis (from Greece), Jordan Burgess (from Newbury, Berkshire) and Pawel Budzianowski (from Poland). All four members met while studying for a master's degree in Machine Learning, Speech and Language Technology at the University of Cambridge. Pawel is moving into a PhD at the university, while the others are hoping to apply themselves to industry.
The final challenge was set by AXA. The challenge dataset contained requests for automobile insurance quotes received by the insurance company from different brokers and comparison websites. The participants were asked to predict whether the person who requested a given quote bought the associated insurance policy. A mammoth 30-hour hackathon witnessed the creation of some of the most innovative and refined solutions to the problem. In the end, teams were ranked according to their prediction score. The performance measure used was log-loss, a measure that strongly penalises predictions that are both confident (probabilities are close to 0 or 1) and wrong.
Max commented: “Our team was delighted and proud to come second after a gruelling 30 hours at the competition. We aimed to come within the top five but we did not expect to come within a hair's breadth of victory over the team from Moscow, which included some of the world's best competitive data scientists and the number one Kaggler.
"We learned that simple models, plied with a bit of feature engineering and intuition, may often outperform the heavy-duty tools. We feel lucky to be part of a sport which enables one to exercise a logical capacity but has so many useful applications in industry, whether that be analysing DNA looking for factors that predispose one to certain diseases, speech recognition or determining whether someone will take up a car insurance quote. We would like to thank the many industry sponsors and mentors who made this event possible, particularly Capgemini who hosted the competition and put forward such excellent mentors.”
The Data Science Game, offering international opportunities for recruitment and leadership recognition, is an association that promotes the development of data science and skills related to scientific challenges addressed to students in computer science, data science, engineering, statistics and/or applied mathematics.
The full press release is available here.
____________________________________________________