Reservoir Computing Sandpit: Funding for Defence & Security

The Engineering and Physical Sciences Research Council (EPSRC), Science and Technology Facilities Council (STFC), and Defence Science and Technology Laboratory (Dstl) have announced a £1,375,000 funding opportunity focused on advancing Reservoir Computing (RC) for national security and defence applications. This initiative, opening November 18, 2025, seeks to develop a deeper understanding of RC – a cutting-edge, in-memory computing technique – and its potential integration into areas like intelligent robotics, artificial intelligence, and 6G networks. Recognizing the impact of rapid technological advances on modern warfare, this funding aims to explore RC’s capabilities where fast prediction, computation, or dynamic control are critical.

Funding Details and Key Dates

This funding opportunity, a four-day sandpit event, is focused on Reservoir Computing (RC) for national security and defence applications. Total funding available is £1,375,000, and the opportunity is led by the Engineering and Physical Sciences Research Council (EPSRC), Science and Technology Facilities Council (STFC), and co-funded by the Defence Science and Technology Laboratory (Dstl). The opening date for applications is November 18, 2025, at 9:00 am UK time, with a closing date of December 16, 2025, at 4:00 pm UK time.

The sandpit aims to generate innovative, multidisciplinary research projects in RC, addressing key challenges identified in collaboration with defence and security stakeholders. Participants will engage in collaborative thinking to construct research projects, culminating in a funding decision made ‘in real-time’ through peer review at the event. The focus is on RC’s potential in areas like intelligent robotics, AI, and dynamic control systems, particularly where low computational burden and model adaption are crucial.

Attendance at the sandpit requires commitment for all online and in-person days, and selected participants must undergo mandatory security checks by Dstl. Researchers new to the defence or security sector are particularly encouraged to apply. While attendance doesn’t guarantee funding, the event fosters collaboration and provides access to stakeholders, aiming to inspire longer-term partnerships and address the UK’s defence and security challenges.

Sandpit Goals for National Security

A £1,375,000 funding opportunity exists for a four-day sandpit focused on Reservoir Computing (RC) for national security and defence. This initiative, led by UK Research and Innovation (UKRI), the Engineering and Physical Sciences Research Council (EPSRC), and in collaboration with the Defence Science and Technology Laboratory (Dstl), aims to generate innovative, multidisciplinary research approaches. The sandpit, opening November 18, 2025, seeks to address challenges and create collaborations between researchers and government stakeholders, with a closing date of December 16, 2025.

Reservoir Computing is a non-traditional computing architecture suited to processing time-series data, potentially benefiting areas like intelligent robotics and cyber-physical systems. Its strength lies in dynamic control, requiring less training data and computational time than traditional Machine Learning. The sandpit will focus on understanding the best combination of RC algorithms, architectures, and substrates for Defence tasks, with a goal of enabling computation in challenging environments, particularly at the edge of networks.

This sandpit seeks to encourage participation from researchers new to the defence and security sector, alongside those with existing links. The process will involve defining research scope, establishing a common language, sharing expertise, and culminating in funding recommendations based on real-time peer review. Participants will have access to government stakeholders to drive research towards real-world scenarios, aiming to inspire longer-term collaborations.

Eligibility Requirements for Applicants

To be eligible for this funding opportunity, applicants must be based at a UK research organisation eligible for Engineering and Physical Sciences Research Council (EPSRC) funding. Participation requires full attendance—both online and in-person—throughout the four-day interactive sandpit. Those selected to attend will also undergo mandatory security checks conducted by the Defence Science and Technology Laboratory (Dstl), and attendance is contingent upon successful clearance.

This funding opportunity is open to organisations with standard eligibility, as defined by EPSRC. The source explicitly states that UKRI-RCN and UKRI-IIASA agreements do not apply, meaning grants cannot include co-leads based in Norway or working with IIASA. International collaborators are permitted, but should be listed as project partners, not co-leads. Uninvited resubmissions of previously submitted projects to UKRI or other funders will not be accepted.

The sandpit aims to foster collaboration between researchers, innovators, and government users, particularly those new to the defence and security sector. Applications are encouraged from individuals with no prior involvement in these areas, alongside those with existing links. The intention is to inspire long-term collaboration and engagement with defence and security stakeholders beyond the sandpit itself, driving research toward real-world scenarios.

International Collaboration Restrictions

This funding opportunity from UKRI and EPSRC, in collaboration with the Defence Science and Technology Laboratory (Dstl), specifically restricts international collaboration. Grants cannot include co-leads based in Norway or at the IIASA institute. All international collaborators, or UK partners from organizations not already approved, must be designated as project partners rather than leads. This ensures the focus remains on UK-based research and security, while still allowing for beneficial international input to the project.

Applicants selected to attend the sandpit will undergo mandatory security checks completed by Dstl. Basic personal information will be required for this process, and attendance at the sandpit is conditional on successfully passing these checks. This highlights the national security focus of the research and the importance of vetting participants who will be exposed to sensitive information and working on defence-related technologies.

The initiative aims to create new collaborations between researchers, innovators, and government users, but actively encourages participation from individuals without prior involvement in the defence or security sector. The source material explicitly states a desire to inspire longer-term collaborations between sandpit attendees and stakeholders, suggesting a strategic effort to broaden the research base in this critical field, while maintaining security protocols.

Resubmission and Prior Funding Policies

Regarding submissions, uninvited resubmissions of projects previously sent to UKRI or any other funder will not be accepted. Applicants should consult EPSRC’s resubmissions policy for further details. This funding opportunity welcomes applications from individuals with or without prior involvement in the defence or security sector, encouraging broad participation and the formation of new collaborations. The intention is for sandpit participants to build lasting relationships with stakeholders in these sectors.

The sandpit process aims to generate innovative research projects through several stages, including defining the scope of research addressing UK defence challenges, establishing a common understanding amongst diverse participants, and facilitating collaborative thinking. Outputs will be captured as highly innovative research projects, followed by a funding decision made through ‘real-time’ peer review at the sandpit.

Reservoir Computing (RC) is described as a non-traditional computing architecture suited for processing time-series data and dynamic control, offering advantages like reduced training time and potential use at the edge of networks. While not fully optimized compared to large deep learning models, RC’s characteristics suggest utility in applications requiring model adaptation, low computational burden, and dynamic control, such as intelligent robotics.

Equality, Diversity and Inclusion Commitment

UK Research and Innovation (UKRI) is committed to equality of opportunity for all funding applicants, actively encouraging applications from a diverse range of researchers. Support is available for individuals needing career breaks, those with caring responsibilities, and for flexible or alternative working patterns. Furthermore, UKRI offers disability and accessibility support throughout the application and assessment process, ensuring inclusivity for all potential applicants seeking funding for the Reservoir Computing sandpit.

The sandpit itself is designed to be an intensive, inclusive, and interactive environment. It aims to bring together a diverse group of researchers from various disciplines and backgrounds within the UKRI’s remit. Recognizing the value of cross-disciplinary collaboration, the event encourages participation from individuals even without prior experience in the defence or security sector, hoping to inspire longer-term partnerships with stakeholders.

This funding opportunity seeks to address critical needs in national security and defence through Reservoir Computing (RC). RC, a type of in-memory computing, shows promise in areas like intelligent robotics and 6G networks, especially where fast prediction or dynamic control is required. The sandpit process will focus on defining research scope, fostering common language, and ultimately, developing innovative projects eligible for funding through ‘real-time’ peer review.

Reservoir Computing: Core Concepts

Reservoir Computing (RC) is a non-traditional computing architecture utilizing a high-dimensional, nonlinear reservoir to perform computational tasks and implement Machine Learning (ML). Importantly, the nonlinearity creates memory of previous inputs, making it well-suited for processing time-series data from dynamic systems, such as those found in intelligent robotics and cyber-physical systems. RC potentially requires less computational time and data for training compared to traditional ML architectures, though optimization remains a challenge.

RC appears particularly relevant for applications where model adaptation, low computational burden, and dynamic control are crucial. The source suggests RC will likely prove most useful at the edge of networks and on small, disconnected devices. Understanding the best combination of RC algorithms, architectures, and substrates for specific Defence tasks is identified as a groundbreaking area for research, potentially enabling computation in challenging physical environments.

A four-day sandpit is being hosted by UK Research and Innovation (UKRI) and the Defence Science and Technology Laboratory (Dstl) to foster innovative multidisciplinary approaches to RC for national security and defence. This collaborative environment aims to generate research applications that account for the needs of UK defence stakeholders, and create new transformative research ideas, with a total fund of £1,375,000 available. The sandpit will focus on identifying key research challenges and fostering collaboration.

Sandpit Structure and Process

The funding opportunity centers around a four-day interactive “sandpit” designed to foster new research into Reservoir Computing (RC) for national security and defence. Participants must be based at a UK research organisation eligible for Engineering and Physical Sciences Research Council (EPSRC) funding and commit to attending all online and in-person sessions. The goal is to generate innovative, multidisciplinary research applications by bringing together researchers, innovators, and government stakeholders – ultimately leading to project pitches for potential funding totaling £1,375,000.

This sandpit process is structured in several stages, beginning with defining the scope of research related to UK defence and security challenges. It then focuses on establishing a common understanding among participants from diverse backgrounds, sharing expertise, and immersing attendees in collaborative thinking. The process culminates in capturing outputs as innovative research projects and making funding recommendations through ‘real-time’ peer review facilitated by a director, mentors, and stakeholders.

Reservoir Computing, the focus of this initiative, is a non-traditional computing architecture particularly well-suited for processing time-series data and dynamic control systems like intelligent robotics. While it requires less computational time and data than traditional Machine Learning, optimisation remains a challenge. The intention is to determine the best combinations of RC algorithms, architectures, and substrates for specific Defence tasks, with a focus on edge computing and small, disconnected devices.