Back to news

Successfully funded projects to pursue innovative applications of AI in research and real world contexts

16 November 2022

The first joint funding call for projects through the Accelerate Science and Cambridge Centre for Data Driven Discovery (C2D3) grant scheme took place this summer, welcoming applications from researchers across the University of Cambridge to start or scale interdisciplinary collaborations in the use of AI for research and innovation. From mapping tree populations in Cambridge to classifying intestinal biopsies, our newly funded projects will pursue applications of AI across a spectrum of research and real world contexts.

Successfully deploying AI to tackle real-world challenges requires effective interdisciplinary collaboration, supported by time and resources to bring together potential research partners, develop new AI tools and software toolkits, and develop new skills or networks. Recognising that this work often falls outside the scope of routine funding calls, the Accelerate and C2D3 funding programme aims to help to fill this gap by offering small grants that can be deployed flexibly. We’re delighted to announce that 2022’s funding scheme will be supporting 9 new projects using AI for scientific discovery.

The funded projects are:

Antimicrobial resistance in farming

Daniel Buhl, PhD Student, Department of Veterinary Medicine

Antimicrobial-resistant bacteria pose a major threat to the healthcare and agricultural system. Tackling this threat requires fast identification of the causes of disease in farm animals, so the right treatment can be deployed in response. Daniel and collaborators have developed a rapid diagnostic method for bacterial identification. The team propose to use a cloud-operated continuous learning system to translate this into a real time output.

Quantifying Design Tradeoffs in Electricity-generation-focused Tokamaks using AI

Hong Ge, Senior Research Fellow, Department of Engineering

Working with collaborators at the University of Cambridge and the UK Atomic Energy Authority, Hong aims to improve the design of Tokamaks (reactors for electricity generation) using Informed Sampling and Bayesian Optimisation and to deliver a useful set of tools that will be assimilated into engineering design systems.

Automated preclinical drug discovery in vivo using pose estimation

Edward Harding, Research Associate and Postdoctoral Neuroscientist, Institute of Metabolic Science, Department of Clinical Biochemistry

This project aims to re-develop and improve the drug discovery pipeline for dementia using computer vision. In this proof-of-principal project, Edward and his collaborators will develop an automated and unbiased validation platform that assesses the effectiveness of different candidate drugs in a mouse model of neurodegeneration.

Causal Methods for Environmental Science Workshop

Sebastian Hickman, PhD student, Department of Chemistry

Sebastian and his collaborators are organising a 2 day workshop in Cambridge to explore how causal AI can help accelerate environmental science. The workshop will be followed by online collaborative sessions to establish a research community, identify problems of mutual interest and to develop new lines of research exploiting recent advances in causal methods to be applied in environmental and climate data science.

Automatic tree mapping in Cambridge

Toby Jackson, Postdoctoral Research Associate, Department of Plant Sciences

Working with the Cambridge Canopy Project, Toby and his collaborators will apply an open source machine learning tool, detecttree2, to provide maps of tree location, size, species and condition across Cambridge. This data will be used by the City Council to aid decision making for the long term management of trees across the city.

Acoustic monitoring for biodiversity conservation

Arik Kershenbaum, Official Fellow and College Lecturer, Girton College and Department of Zoology

Monitoring and tracking endangered species is an essential part of conservation efforts. Acoustic monitoring, using an animal’s call to track their location, is a non-invasive and effective way of collecting this data. This project will convene a 5-day workshop bringing together biologists and computer scientists to generate new impetus and collaborations in applying AI techniques to detect and classify animal calls automatically, to support monitoring of animals in biodiversity conservation.

AI, mathematics and string theory

Challenger Mishra, Departmental Early Career Academic Fellow and Affiliated Lecturer, Accelerate Science, Department of Computer Science and Technology

Challenger and his collaborators will develop interactions between machine learning researchers, and pure and applied mathematicians through workshops focussing on string theory and machine learning and AI for Mathematical Discovery.

Theoretical, Scientific, and Philosophical Perspectives on Biological Understanding in the age of Artificial Intelligence

Srijit Seal, PhD Student, Department of Chemistry

While machine learning methodologies have seen widespread adoption in modern biology, the scientific rationale for applying these models remains underdeveloped. This project will bring together scholars from different backgrounds including Machine Learning scientists, biologists and philosophers to discuss how we can plan novel hypotheses from data and if Artificial Intelligence is the suitable language for describing biological systems.

AI in pathology: optimising a classifier for digital images of duodenal biopsies

Elizabeth Soilleux, Associate Professor/Research Group Leader and Honorary Consultant Pathologist, Department of Pathology

How can AI help more effectively diagnose coeliac disease? Elizabeth and her collaborators have developed an algorithm that can identify coeliac disease from duodenal biopsies, with 97–99% accuracy. With this funding, the team will spend 6 months testing and “productising” their algorithm, embedding it in a software format that can run on a pathologist’s computer, ready for a multicentre clinical trial of diagnostic accuracy and user acceptability.

Successful projects will take place over the coming months, with researchers sharing updates through the Accelerate Programme blog and engaging with the AI community at Cambridge. The wide range of projects supported through this call will catalyse further discoveries to accelerate scientific progress and create AI tools that are capable of delivering benefits for science and society.