Context: Invasive insect species have the potential to outcompete or predate native species and bring disease. As mobile devices increasingly support biodiversity monitoring, acoustic detection and identification of insects opens up a new avenue to expand the coverage of biodiversity monitoring in the United Kingdom. Such technology is ideally suited for surveillance of invasive species, where the species density is initially low, meaning surveillance effort can be costly and uncertain but still has to be balanced against the potential economic cost of successful invasion. Challenge: A major challenge, in tandem with the great potential offered by mobile acoustic sensing, is the gap between the large volume of raw acoustic recordings generated through passive monitoring and the data processing capacity necessary to promptly extract valuable information from audio contents. Moreover, positive detections must be relayed swiftly to alert research communities, the wider public, and policy makers. Delays can grow excessively in a nationwide mobile acoustic monitoring programme where thousands of mobile devices record sound events as they monitor for invaders. Failing to address this challenge will limit the adoption of a mobile acoustic system for monitoring of invasive species and constrain our ability to deliver timely insights into habitat connectivity and species mobility based on acoustic sensing.

The UK's 2022 National AI Strategy recognizes the potential of AI for both growing prosperity and human flourishing, but AI's impact is hindered by a core obstacle-skills shortages in end-to-end autonomous systems, the systems, like autonomous robots, required to integrate AI within real-world settings. The Autonomous Intelligent Machines and Systems (AIMS) Centre for Doctoral Training (CDT) aims to tackle this obstacle by training cohorts in theoretical and systems skills in autonomous systems, with industrial partnerships co-creating AIMS training and ensuring the delivery of Oxford's research to various UK sectors, including transport (partnering with Toyota, Oxa, Waymo), extreme environments (Satellite Catapult, Trillium, Marine AI) and life-critical decisions (Aioi Nissay Dowa Insurance, Mind Foundry). AIMS is building on success in training future leaders in autonomous systems. Since its inception in 2014, AIMS has delivered high-quality research and impact, with students publishing 157 papers in high-impact venues such as Science, Nature Communications, including 43 papers at NeurIPS, arguably the premier venue in AI. AIMS students have won best paper awards from CVPR, outstanding reviewer awards from ICLR and NeurIPS, awards from Qualcomm, and fellowships from IBM. Two AIMS students published a sequence of papers on the effectiveness of interventions against Covid-19 that led to broad media and policy impact, along with an Impact Award from the MPLS division of the University of Oxford. AIMS graduates have secured posts in top universities, companies, and have founded their start-ups. Indicative of AIMS's success is that in its most recent five years, AIMS averaged 223 applications per year for only 12 places. In the first year, our cohort-focussed programme trains students in 18 bespoke week-long graduate courses, centred in Machine Learning, with spokes in Robotics and Vision, Control and Verification, and Cyber-Physical Systems. Training also includes components devoted to transferable skills (e.g. entrepreneurship), responsible research and innovation and AI safety and governance. Subsequently, students complete two 10-12 week mini-projects, followed by a three year research project, most designed in partnership with industry. AIMS is also building on success in engagement with industry, with partners having contributed £2.8 million of cash support in the last five years of AIMS and promising another £2.8 million in cash support to this proposal. In addition to this direct support, our partners have been generous with in-kind support-Nvidia, YouGov, Samsung, and Mathworks have successfully organized workshops that foster co-creation of research problems in the context of real-world applications and in training on specific technical tools (e.g. Nvidia's CUDA toolkit). The renewal of our funding will permit AIMS to enhance a developing cluster addressing life-critical autonomous systems, linked to new partnerships in healthcare and insurance. AIMS also intends to expand into the related area of biotechnology, involving work on robotic bioreactors (see Steel's prize-winning Chi.Bio Bioreactor) and the automation of biological design. Finally, AIMS aims to explore AI applications in ecology and environment, aligning with existing projects like the Embodied Intelligence Programme Grant EP/V000748/1. AIMS's chosen primary focus area is 'delivering an EPSRC research priority', and AIMS is aligned with EPSRC's mission-inspired research priority of 'artificial intelligence, digitalisation and data: driving value and security'.

Artificial intelligence (AI) is on the verge of widespread deployment in ways that will impact our everyday lives. It might do so in the form of self-driving cars or of navigation systems optimising routes on the basis of real-time traffic information. It might do so through smart homes, in which usage of high-power devices is timed intelligently based on real- time forecasts of renewable generation. It might do so by automatically coordinating emergency vehicles in the event of a major incident, natural or man-made, or by coordinating swarms of small robots collectively engaged in some task, such as search-and-rescue. Much of the research on AI to date has focused on optimising the performance of a single agent carrying out a single well-specified task. There has been little work so far on emergent properties of systems in which large numbers of such agents are deployed, and the resulting interactions. Such interactions could end up disturbing the environments for which the agents have been optimised. For instance, if a large number of self-driving cars simultaneously choose the same route based on real-time information, it could overload roads on that route. If a large number of smart homes simultaneously switch devices on in response to an increase in wind energy generation, it could destabilise the power grid. If a large number of stock-trading algorithmic agents respond similarly to new information, it could destabilise financial markets. Thus, the emergent effects of interactions between autonomous agents inevitably modify their operating environment, raising significant concerns about the predictability and robustness of critical infrastructure networks. At the same time, they offer the prospect of optimising distributed AI systems to take advantage of cooperation, information sharing, and collective learning. The key future challenge is therefore to design distributed systems of interacting AIs that can exploit synergies in collective behaviour, while being resilient to unwanted emergent effects. Biological evolution has addressed many such challenges, with social insects such as ants and bees being an example of highly complex and well-adapted responses emerging at the colony level from the actions of very simple individual agents! The goal of this project is to develop the mathematical foundations for understanding and exploiting the emergent features of complex systems composed of relatively simple agents. While there has already been considerable research on such problems, the novelty of this project is in the use of information theory to study fundamental mathematical limits on learning and optimisation in such systems. Information theory is a branch of mathematics that is ideally suited to address such questions. Insights from this study will be used to inform the development of new algorithms for artificial agents operating in environments composed of large numbers of interacting agents. The project will bring together mathematicians working in information theory, network science and complex systems with engineers and computer scientists working on machine learning, AI and robotics. The aim goal is to translate theoretical insights into algorithms that are deployed onreal world applications real systems; lessons learned from deploying and testing the algorithms in interacting systems will be used to refine models and algorithms in a virtuous circle.

Artificial intelligence (AI) is undergoing an era of explosive growth. With increasingly capable AI agents such as chatGPT, AlphaFold, Gato and DALL-E capturing the public imagination, the potential impact of AI on modern society is becoming ever clearer for all to see. APRIL is a project that seeks to bring the benefits of AI to the electronics industry of the UK. Specifically, we aspire developing AI tools for cutting development times for everything from new, fundamental materials for electronic devices to complicated microchip designs and system architectures, leading to faster, cheaper, greener and overall, more power-efficient electronics. Imagine a future where extremely complex and intricate material structures, far more complex than what a human could design alone, are optimised by powerful algorithms (such as an AlphaFold for semiconductor materials). Or consider intelligent machines with domain-specialist knowledge (think of a Gato-like system trained on exactly the right milieu of skills) experimenting day and night with manufacturing techniques to build the perfect electronic components. Or yet what if we had algorithms trained to design circuits by interacting with an engineer in natural language (like a chatGPT with specialist knowledge)? Similar comments could be made about systems that would take care of the most tedious bits of testing and verifying increasingly complex systems such as mobile phone chipsets or aircraft avionics software, or indeed for modelling and simulating electronics (both potentially achievable by using semi-automated AI coders such as Google's "PaLM" model). This is precisely the cocktail of technologies that APRIL seeks to develop. In this future, AI - with its capabilities of finding relevant information, performing simple tasks when instructed to do so and its incredible speed - would operate under the supervision of experienced engineers for assisting them in creating electronics suited to an ever-increasing palette of requirements, from low-power systems to chips manufactured to be recyclable to ultra-secure systems for handling the most sensitive and private data. To achieve this, APRIL brings together a large consortium of universities, industry and government bodies, working together to develop: i) the new technologies of the future, ii) the tools that will make these technologies a reality and very importantly, iii) the people with the necessary skills (for building as well as using such new tools) to ensure that the UK remains a capable and technologically advanced player in the global electronics industry.