Data scientists from Imperial and London School of Economics launched a new seminar series to jointly tackle their unsolved research problems.
On 30 September 2022, the Data Science Institute at the London School of Economics hosted the first seminar for a new series that forms part of an ongoing bilateral collaboration with Imperial College London.
This new series of seminars aims to foster innovations by bridging the gap between social sciences, computer sciences and STEM subjects through presenting unsolved problems and crowdsourcing solutions from experts across these fields.
The series avoids the classic seminar format of long presentations and restricted chance for audience contributions by limiting the presenter to just an informal twenty-minute outline of their research.
In this first seminar Dr Rossella Arcucci presented her unsolved problem relating to data science and machine learning for climate models.
Unsolved problems in data assimilation and AI for climate models
Data assimilation, or the process of keeping a model on track by constantly correcting it with fresh observations, and machine learning as a way of accurately predicting things in real-time can help save lives, from predicting natural disasters to diagnosing health-related problems.
Dr Arcucci’s unsolved problem in the field of data assimilation and machine learning involves the accurate and efficient predictions of climate-related problems using AI models based on artificial neural networks. For instance, there is a lack of data when it comes to the climate, and currently Dr Arcucci’s work uses generative AI models to create synthetic realistic data. Although this process works well with some applications, there are still some unanswered questions on how well these predictions match up with real-world observations about the climate.
Aside from this, her research offers further unsolved problems when it comes to minimizing the ‘noise’, or errors, in datasets, automatic fine-tuning of the systems to incorporate the most up-to-date data, and improving compression models to ensure that information is not lost when big data is compressed to be used.
The ‘DSI Squared’ initiative joins the Data Science Institutes from both Imperial College London and the London School of Economics (LSE).
When it comes to data science research and its impact, the LSE’s strengths in the social sciences naturally complements Imperial’s strengths in science, technology, and medicine.
By working together, the team hopes this initiative will enhance their joint influence on policy in wide scope domains – areas where alternative facts compete with scientific findings for influence in the policymaking process.
To keep up to date with the latest DSI Squared events and news, sign up to the Imperial DSI Newsletter.
Article text (excluding photos or graphics) © Imperial College London.
Photos and graphics subject to third party copyright used with permission or © Imperial College London.