Recipient organisationScience and Technology Facilities Council
Funding£125K
PeriodSept 2025 — Mar 2026
In plain English
AI plain-English summary
Diamond Light Source, a UK national synchrotron, will soon use artificial intelligence to decide in real time which samples to analyse next, bypassing the current bottleneck where data piles up faster than humans can interpret it. The problem is straightforward: UK physical sciences facilities generate enormous volumes of experimental data, but the tools for analysing that data and steering experiments have not kept pace. Researchers often run experiments blind, only discovering later that they missed crucial measurements or wasted time on uninformative samples. This project aims to close that loop by embedding AI directly into the experimental workflow. If successful, the approach will transform how the UK’s national research facilities operate. Instead of scientists manually inspecting each diffraction pattern—a technique used across materials science, chemistry, and structural biology—the AI will flag interesting results, adjust parameters, and guide the next measurement autonomously. This could dramatically accelerate materials discovery, drug development, and battery research. The project also builds a curated, AI-ready repository of diffraction data, ensuring reproducibility and enabling future machine-learning models to extract deeper insights from past experiments. The impact is on the infrastructure of science itself: faster, smarter, and more efficient experimental campaigns.
View original technical description
The Ada Lovelace Centre (ALC) is leading a transformative initiative to accelerate experimental science through AI-driven high-throughput screening and autonomous experimentation. This project addresses a growing challenge: while the volume and complexity of experimental data from UK physical sciences facilities is rapidly increasing, the tools for experimentation, analysis and interpretation have not kept pace. This project utilises the expertise of ALC, Diamond and PSDI to tackle both the application and data challenges associated with transforming experimentation and analysis through AI. ALC is spearheading the development of AI applications that impact facility science by enabling rapid, predictive analysis of experimental data, with a focus on real-world use cases and facility integration. Diamond is a core large scale facility supporting the broad science needs of the UK science community. PSDI complements this by building robust, AI-ready data pipelines and a prototype repository, ensuring that curated, high-quality datasets are available to power these applications. Diffraction has been selected as the primary use case due to its broad relevance across central facilities (e.g. Diamond Light Source, ISIS), lab facilities such as those doing materials discovery, and EPSRC National Research Facilities (NRFs). It also offers a rich landscape of data sources, enabling the development of generalisable AI models and workflows. This cross-cutting set makes an ideal testbed for demonstrating the value of AI in experimental science and ensuring wide impact. This will be achieved: · Enabling high-throughput screening to guide experimental decision-making in near real-time. · Implementation of AI-driven feedback loops which implement decision making or improve parameter selection. · Deliver curated diffraction datasets with well-defined outcomes, working with PSDI, to support reproducibility and model training. · Apply AI analysis to the larger curated datasets to demonstrate extraction of additional meaning By pairing ALC’s application-driven innovation, PSDI’s data infrastructure expertise and Diamond experimental and data challenges, this project will draw together domain experts necessary for the collection of highly curated and trustworthy datasets with experimental and application focused experts. This will deliver scalable, interoperable solutions that enhance the efficiency and impact of the UK’s national research facilities.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know