Led by Dr Andrew Duncan, fundamental research in data science and artificial intelligence is a key capability in our Turing.2.0 strategy that will support the delivery of our three strategic goals.
The Turing's approach is to focus on end-to-end interdisciplinary pathways which can take novel ideas in fundamental research, through applied research identified in our grand challenges (environment and sustainability, health, defence and national security), to end-user impact at scale.
The primary aim of our programme is to democratise data science and AI through the development of new tools, methods and theory which:
- Enable the application of AI methodology across new domains, as prioritised by the Turing's grand challenges.
- Support the democratisation of AI development and access through fundamentally new approaches.
- Create tools and practices for sustainable AI, focusing on developing and implementing AI technologies that reduce environmental impact and enhance long-term sustainability.
- Develop the theory and principles of safe and trustworthy AI, embedding these principles into the work of the Turing and disseminating them to international AI research and practitioner communities.
We are aligning our portfolio of activities into defined missions which will tackle targeted problems. Our initial mission in development is AI for physical systems, with the aim of developing the next generation of foundational methods, tools and theory to enable modelling, prediction and control of physical systems.
To achieve this, we are creating a multi-disciplinary, mission-driven team which will collaborate with national and international centres of excellence. Initially, we will be focusing on three strands:
Strand 1: Probabilistic and generative models for modelling and prediction of physical systems.
Strand 2: Bridging the divide between data-driven and mechanistic models.
Strand 3: Accelerating large-scale computational simulations through machine learning.
