The Computer Science day brought 200+ students, sponsors and researchers across Oxford and Cambridge, our biggest event yet.

It was a wonderful hub of discussion; we attracted a vast range of students, from first-year undergrads to post-doctoral researchers, and it was energising to see people from so many backgrounds sharing their work and interests with each other. There was not a moment the auditorium wasn’t buzzing, and the happy feedback we got certainly reflected the successes of the day.

We invited three keynote speakers from Google DeepMind, Imperial and TPO securities to talk about their work. I particularly enjoyed hearing about Eleonora Giunchiglia’s talk about how requirements can make machine learning safer- using loss function modelling and logic similar to techniques I am currently revising for in my finals modules, it is possible to reduce the ‘shortcuts’ that machine learning modules create when given certain safety constraints, so that they produce output that reflects an intelligent, reasoning being more closely. These shortcuts often skip the normal reasoning steps a human would take when faced with a constraint, which result in surprising behavioural differences we see in LLM outputs. For example, classifying what is obviously a pig (to humans) as a dog, because the LLM only looks at ears and fails to analyse perhaps the pig’s snout too. This motivates the goal for more reliable, safer machine learning models that wouldn’t make simple mistakes that we wouldn’t expect a human to make, due to shortcuts in their reasoning steps.

Alongside the keynotes, the day allowed over 40 students to present posters and orals celebrating their work. We had a workshop run by Immersive Labs, one of our sponsors, on a cyber crisis simulation, and sponsor stalls dotted throughout the venue with employees on hand to chat to students about working at their companies. We facilitated 1-1s for handpicked students and sponsor representatives to foster interesting conversations and communication.

The day culminated in a wonderful formal dinner at Trinity College, before the Cambridge students begun their long coach drive back home. Computer Science currently has one of the worst gender ratios for undergraduates across the courses available at Oxford, and I am really passionate about changing this. I am so proud of my co-organisers and myself for fostering an environment where women in Computer Science truly shone and inspired each other. It was an honour to carry on this legacy, and I hope it inspires more women to join the field.