About DISI

The Centre for Data Intensive Science and Industry (DISI), is hosted within the Department of Physics and Astronomy.

It is a hub that plays a pivotal role in engaging PhD and MSc students in the realms of AI training, research, and the exchange of knowledge between academia and industry. The Provost’s Strategic Fund has supported DISI by providing 2 Lectureships, a Centre Manager and a 1-year Postdoc. This has enabled us to create an umbrella structure, DISI, which now incorporates:

The CDT-DIS, which was established in 2017, and selected recently by STFC for an extension for 3 more cohorts, hence operating until at least 2028, with typically 10 PhD students per year.
An MSc 1-yr programme in Scientific and Data-Intensive Computing, with typically 70 students per year.
The relationships with over 30 industry partners (IPs) are a vital component of DISI's mission to foster collaboration and knowledge exchange between industry and academia in the fields of Machine Learning and Artificial Intelligence. The placement opportunities provided by these IPs are an integral part of the training and development of CDT DIS students.

Vision

Train the next generation of leaders in DIS for roles in academic and industry.
Exchange knowledge between fields and academia/industry.
Develop multi-disciplinary collaboration in DIS (cross-departmental, nationally, internationally, between academia/industry).
Foster responsible research and innovation.

Examples of current and recent DISI-related activities

A mini-workshop was organised on 20th April 2023, with colleagues from Google-DeepMind on cutting-edge applications of graph neural networks (GNNs) in the modelling of complex physical systems. The mini-workshop was preceded by a public seminar and attracted over 150 attendees.
DISI academics have been heavily involved in national capacity building and networking activities with the Turing. As a result, two Turing Interest Groups have been set up, one in Space Science and one on Simulation-Based Science.
On the international front, we completed in March 2023 the Newton-funded programme with Jordan, that trained successfully about 30 Jordanian students in DIS, including 2 online courses and a summer school held at UCL in July 2022.
We have established links with the Flatiron Institute (in NYC) and with the President of the Simons Foundation, who visited UCL on 27 June 2023. We held that day a roundtable meeting, with presentations by post-docs on AI at UCL, from brain research to fundamental physics.
Members of DISI have recently led a consortium of colleagues from six UCL departments that was successful in being selected through the UCL internal selection process as one of only two proposals to be submitted in July to UKRI in the funding call for applied AI CDTs.. The title of the proposed CDT is “UKRI AI CDT for Data Intensive Physical Systems” and focuses on Fusion, Climate and Fundamental Physics.

Our planned further DISI activities:

To run 1-2 topical workshops per year, with participants from both academia and industry.
To encourage candidates for advance fellowships (e.g. ERF, URF, FLF and similar international schemes) to have UCL (and the DISI Centre) as their host.
To establish a Visiting Fellows scheme, hosting DIS experts from other universities and industry for periods of 1-12 months.
To establish a network with half a dozen ML/AI centres around the world, for supporting closer collaborations, e.g. exchange schemes, workshops (in person and online) etc.
To enhance the cooperation with the other 4 STFC-funded CDT-DIS centres and with the Alan Turing Institute.
To expand the ML research directions beyond Astrophysics and HEP, e.g. towards Fusion and Climate Change.