APEX Awards 2023

Dr Mazviita Chirimuuta

University of Edinburgh

Co-applicant(s): Dr Juan Manuel Parrilla Gutierrez, Glasgow Caledonian University

APX\R1\231098

£91,900.37

Does cognition predate life? A philosophy-inspired engineering approach to defining and testing proto-cognitive processes in inert matter

The documentary “Encounters at the End of the World” follows a penguin in Antarctica. This penguin first appears following another group who are going from the main colony to the feeding grounds. Halfway through the penguin stops and instead goes alone towards the mountains, several miles away, towards certain death. Did the penguin decide to do this, or did the action result from an unusual fluctuation in the animal’s brain chemistry? Decision making in animals evolved in order to help them survive, but are all thought processes the result of evolution, or is there a chemical foundation to cognition predating the origin of life? This is the question we explore in this proposal.

The emerging field of Basal Cognition investigates seemingly intelligent behaviours, like decision making and adaptation in the simplest life forms such as bacteria. Complex, non-living chemical systems display some of these behaviours. Is it meaningful to say that they have an elementary form of intelligence, which we can call “proto-cognition”? To answer this, the proposal combines the forces of philosophy and engineering. The philosophical work involves defining proto-cognition and proposing ways that it could be measured experimentally. On the engineering side, we will develop Artificial Intelligence models informed by this definition to simulate proto-cognition in inert matter. We hope to use AI to understand how cognition could have emerged from basic chemical principles. If successful, this research will help to answer one of the most intriguing scientific and philosophical questions: how does mind arise from non-thinking matter?

Professor Robin Coningham

Durham University

Co-applicant(s): Professor David Toll, Durham University; Dr Mohammed Seaid, Durham University

APX\R1\231178

£96,328.44

Exploring the feasibility of regenerating Medieval Licchavi Period irrigation infrastructure in the Kathmandu Valley, Nepal

Witnessing a 1362.5% population increase to 1.5 million from the 1950s, the Kathmandu Valley contains one of the world’s fastest growing urban populations, and one which is predicted to increase to 2.2 million by 2035 (UN 2022). This dramatic increase not only exposes more households to the risks of natural disasters within this seismically prone region, as demonstrated by the 2015 Earthquake, but also places pressure on essential services, which have not kept pace with increasing demand. Access to clean and safe water supplies for domestic and agricultural use is a potential source of conflict and insecurity within the Valley, with unchecked development leading to demand outstripping supply, as

well as damaging and straining current water capacity. Solutions, such as tanking water to areas in need is unsustainable; however, there is potential that lessons learned from past historic infrastructure can provide alternative and sustainable solutions to this chronic situation.

Recent post-earthquake fieldwork at damaged heritage sites within Kathmandu undertaken by archaeologists, geoarchaeologists, philologists, architects, engineers, artisans, and local communities revealed historic indigenous technologies with climatic and seismic adaptations that have been successfully refitted into sustainable reconstructions with continued community monitoring and management. We propose to expand our network to investigate Licchavi Period (5th-8th centuries CE) irrigation systems, recorded within inscriptions and located archaeologically, and explore the potential regeneration of historic irrigation systems for sustainable water supplies managed by resilient communities to mitigate continued threats from increased population, rapid urban development, seismic events, flooding and climate change.

Professor Timothy Heaton

University of Leeds

Co-applicant(s): Professor Paula Jo Reimer, Queen's University Belfast

APX\R1\231120

£88,315

Understanding and Explaining Variations in the Rate of Radiocarbon Samples

Explaining how the environment around us affects our lives is one of the biggest questions we face today. A key approach that can help is to study the records of our past. The last 55,000 years have seen pivotal changes in many aspects of society, as well as substantial changes in our environment and climate. What might have caused these societal and population changes, and are they linked to the changes seen in the surrounding environment? How does the frequency of fire, or floods, change with environmental factors? Have abrupt climatic changes led to large human migrations? Can variations in temperature, solar activity, or rainfall cause societal collapse?

Answering such questions is made more challenging since the timescales on which most archaeological records are created rely upon radiocarbon. All radiocarbon measurements need to be calibrated to be understood on a calendar age scale. The resultant calendar age estimates are complex and uncertain. This makes assessment of causal and explanatory relationships between archaeological proxies for the rate of extreme events (or population size) and those obtained by other means difficult. Existing methods of statistical regression are not appropriate.

This project will develop novel methods to test whether variations in the frequency of extreme events or the size of a population (inferred by the calendar rate of samples preserved in the archaeological record) can be explained by proposed climatic and environmental factors. This work can only be achieved through the successful combination of statistical, archaeological, and geoscientific knowledge provided by our team.

Dr Lina Jansson

University of Nottingham

Co-applicant(s): Dr Silke Weinfurtner, University of Nottingham; Dr. Marco Iglesias, University of Nottingham

APX\R1\231030

£90,895.24

Bringing the Cosmos to the Lab: Explaining via Analogue Gravity Quantum Simulators

Our team seeks to understand how we can learn about distant inaccessible systems—such as black holes or events in the early universe—by studying systems in the lab that can replicate some of the behaviour of these systems. We would like to use lab-based simulators to extend our knowledge beyond what we can calculate and to transfer that knowledge to inform our understanding of black holes and the early universe. However, simulators are not exact replicas of the systems that we are interested in and we need to confront the question of how reliable the inferences drawn from our simulators are and what evidence we can bring to bear to increase their reliability.

Since we cannot test our simulators against observation from inaccessible systems in a straightforward way, we confront philosophical questions about the role of criteria such as simplicity, similarity, and explanatory power in science when reasoning from simulators to distant systems. This project brings together philosophical work on the methodology of science and scientific work on gravity simulators to determine the extent to which we can trust conclusions drawn from simulators.

One of the most exiting aspects of this work is that it brings together philosophy and science in a way that used to be seen in early modern science but that is unusual in contemporary work. The question that we tackle cannot be answered by either discipline alone and shows that integrated science and philosophy of the kind historically seen in natural philosophy remains relevant today.

Dr Vitaliy Kurlin

University of Liverpool

Co-applicant(s): Professor Andrew Cooper, University of Liverpool

APX\R1\231152

£99,999.39

New geometric methods for mapping the space of periodic crystals

In 1869, Mendeleev arranged the known chemical elements into a spatial map – the periodic table - which grouped them according to their properties. The appearance of ‘gaps’ in the table spurred the discovery of new elements, while the attempt to understand the physical rationale behind its structure drove revolutionary advances in both physics and chemistry.

We aim to extend Mendeleev’s idea to all periodic structures. Our approach uses geometry - the spatial relationships between atomic nuclei in the structure – as a classification tool.

This approach operates at the interface of materials science and mathematics, in the remits of the Royal Academy of Engineering and the Royal Society, respectively – satisfying thoroughly the objectives of the APEX scheme.

We have recently developed the first unambiguous and stable-under-noise representation of crystal structures as a finite collection of real numbers. We proved that this is sufficient to uniquely represent any structure, and that conversely any such finite collection of numbers can be uniquely reconstructed as a potential crystal.

This representation enables us to situate structures in a geographical space just as Mendeleev situated the elements in his table. Reconstruction means that we can not only see the gaps in this crystal map but determine the structures that should exist there.

This map has the potential for changing paradigms in materials design. The hunt for materials appropriate for energy storage, catalysis or carbon capture, currently driven by expensive trial-and-error methods, could instead begin by simply selecting the relevant area of the map.

Dr Hemma Philamore

University of Bristol

Co-applicant(s): Dr Emma Liu, University College London; Dr Iwona Gajda, University of the West of England; Dr Jiseon You, University of the West of England

APX\R1\231103

£99,232.05

Robots in the Wild – Environmentally trustworthy robots for monitoring challenging natural ecosystems

In order to tackle the climate crisis and mitigate the impacts of natural hazards, understanding and monitoring our environment is essential. Robots and autonomous systems are increasingly being used as essential data-gathering instruments and must be trustworthy in terms of functionality, environmental impact and social acceptability. As a multi-disciplinary team of engineers, environmental researchers and monitoring organisations, we will, first, co-develop recommended best practices for deploying autonomous with net-zero environmental impact using the volcanic environment as a challenging exemplary case study ecosystem. Second, recent innovations in bioelectrical systems and soft and bio-hybrid robots will be translated into robotic hardware demonstrators that serve a positive ecological function. Finally, we will assess the resilience, degradability and social acceptability of the demonstrators using a combination of quantitative experimental testing and evaluation against the proposed best practice recommendations.

Dr Neil Stephens

University of Birmingham

Co-applicant(s): Dr Mariana Petronela Hanga, University College London; Dr Mariela de Amstalden, University of Birmingham

APX\R1\231024

£95,320

Integrated risk mapping of culture meat

Cultured meat is a technology that seeks to grow muscle in vats for consumption as meat. It is a fairly new technology, but has been available to buy in small quantities in Singapore since 2020, probably in the USA in 2023, and the UK following after. The idea is to produce meat that is better for the environment, better for health, and better for animals. However, lots of uncertainties remain about exactly how it will be produced at scale, what the impacts on society will be, and how it should be regulated. In this project we bring together the unique combination of a chemical engineer who works on producing cultured meat, with a sociologist who works on the politics of cultured meat, and a lawyer who works on anticipating what the regulations might be. Through discussions between us, literature reviews, and engaging with other experts, we will produce an analysis that integrates all our insights about the future of cultured meat. We will do this by focusing upon the ‘risks’ of cultured meat, to capture all the things that might become negative outcomes, including

technical elements of the process that might go wrong (like the energy needs or cost), social impacts (like changes to jobs or inequalities), and legal impacts (like moving it between countries and making it safe). Importantly, we focus on how the technical, social, and legal aspects influence each other. We will display our findings in a visual map and share it with scientists, policy-makers and publics.

Dr Ozlem Ulgen

University of Nottingham

Co-applicant(s): Dr Pavel Naumov, University of Southampton

APX\R1\231097

£99,985.13

Responsibility for human-AI split-second decision-making (ROUTE)

Drawing on our expertise in law, ethics and regulation of AI, and mathematical logic and game theory, we seek to understand how human-AI split-second decisions are made, and how this may inform attribution of legal responsibility to humans. Examples of split-second decisions include automated stock trading, self-driving cars, and autonomous weapon systems. We focus on these types of decisions because they have many human-AI interaction points, require quick responses in pressurised circumstances, and have potentially harmful consequences (e.g., financial loss, road traffic accident, injury and fatality).

An exciting and innovative aspect of our project is combining law and science to explore the possibility of assigning legal responsibility. If there are identifiable routes to human-AI split-second decision-making, it becomes easier to assign responsibility for when things go wrong or harmful consequences follow. Establishing responsibility in this way will promote transparency and accountability for AI systems, and guide public and private sector organisations using human-AI decision-making. It will also enable public confidence in the use of human-AI decision-making and benefit society by identifying who is responsible for decisions when things go wrong.

We use the legal concept of responsibility, mathematical logic, and game theory to explore human-AI split-second decision-making. We identify when and how humans are involved, and the extent of their involvement. To help us assign responsibility, we use the PI’s “human-centric and lifecycle approach” focusing on human involvement at different stages in the AI system lifecycle.