MINDS CDT 3rd Annual Symposium
How will your research affect people and the economy in the future? MINDS CDT students posed this important question as part of their annual symposium which this time focused on impact and responsible research.
The 3rd MINDS CDT Symposium took place on 24th January 2025 at the Leonardo Southampton Hotel with around 40 attendees. As in previous years, students of the CDT in Machine Intelligence for Nano-Electronic Devices and Systems (MINDS) showcased their research and competed for two prizes for the best presentation.
Beside the MINDS CDT students, the event organisers welcomed the first cohort of a new UKRI CDT in AI for Sustainability (SustAI).
As with the previous events, students were asked to deliver a 10-minute presentation on their research followed by questions from the attendees. This year, however, the PhD students were asked to focus their presentations on responsible research and impact. To enhance this, students were asked to use the RRI 4P AREA framework, a tool used in Responsible Research and Innovation (RRI) that helps researchers understand if their work is ethical and in line with social norms and values. Including the 4 Ps (Purpose, Product, People and Process) the students elaborated on the potential benefits and drawbacks of their research.
After hearing 14 great presentations, the attendees had to complete the challenging task of ranking the talks and finally two winners were agreed. The members of the management teams of both MINDS CDT and SustAI CDT awarded the prize for best presentation to Yaseen Mohammed Osman. The student of cohort 5 impressed the judges with his presentation on ?Low-Memory and Fast Tuning of Large Language Models for Dynamically Changing Tasks?. His presentation focused on reducing the environmental impact of computing resources by tailoring LLMs through fast-tuning.
Oliver Grainge, a student of cohort 4, received the audience?s best presentation award for his talk on ?Harnessing Sparsity in Visual Geo-localization?. Oliver emphasised the potential of neural network pruning to optimise Visual Place Recognition leading to substantial gains in memory efficiency while maintaining high accuracy. All the students showed great commitment to the theme of their presentations demonstrating they are not only advancing their technical expertise but also fostering a culture of ethical innovation. Prizes of £50 Amazon vouchers were given to the 2 winners by guest judge and director of SustAI CDT Professor Enrico Gerding.