We are excited to share that our MRes student Mary-Brenda Akoda (I-X and Department of Computing) has been awarded the Fetch Compute Grant. The grant is part of I-X Business Partner’s Fetch.ai Innovation Lab’s initiative, which aims to revolutionise AI infrastructure by providing free access to high-performance GPUs for AI researchers, developers, and open-source innovators. Through this grant, users can access advanced computing resources, including high-end Nvidia GPUs designed to support advanced AI training and computation-heavy projects.

Mary-Brenda’s project will employ Deep Generative AI to accelerate MRI reconstruction, offering quicker and higher-quality imaging. MRI reconstruction, which involves MR images from raw frequency data, is an inherently challenging and ill-posed inverse problem. Traditional methods often face issues such as noise sensitivity, dependence on hand-crafted priors, and computational inefficiencies. Deep learning-based approaches have shown promise by utilising large datasets, but they still struggle with adapting to variable undersampling scenarios. Additionally, their slow reconstruction speeds hinder clinical feasibility. This project aims to address these challenges by developing new deep generative modelling frameworks that reduce reconstruction times while maintaining high-quality MR imaging.

Mary-Brenda’s project is one of the first three projects awarded the Fetch Compute grant. The additional computational resources provided by the grant will be instrumental in accelerating the training of more advanced machine-learning models. This enhanced capability will help push the boundaries of state-of-the-art generative models in clinical applications. In the long run, the goal of the project is to transform medical imaging, making diagnostics faster, more reliable, and widely accessible.

Commenting on the importance of the award, Mary-Brenda said: “Receiving the Fetch.ai Compute Award marks an exciting milestone in my journey to push the boundaries of what’s possible in AI for Healthcare. It provides me with more freedom to explore and implement innovative solutions that could dramatically improve diagnostic imaging, reduce patient wait times, and ultimately enhance patient outcomes worldwide.”

Dr Chen Qin (I-X & Department of Electrical and Electronic Engineering), the research group lead and Mary-Brenda’s supervisor, said: “It is fantastic that Mary Brenda-Akoda has been awarded the Fetch Compute Grant. This grant will significantly accelerate her research on generative models in medical imaging. With access to more GPU power, she will be able to train more advanced machine-learning models, pushing our research on cutting-edge generative models for clinical applications even further.“

Speaking about the vision behind Fetch Compute, Sana Wajid, Senior Vice President of Fetch.ai Innovation Lab, said: “Fetch Compute is more than just a program – it’s about making innovation accessible. My vision is to bring cutting-edge AI research and innovation to people’s doorstep and by strengthening our collaboration with I-X, we’re creating an ecosystem where researchers and developers can push boundaries of what’s possible without limitations. Fetch Compute is part of our bigger strategic goal for driving innovation through our innovation labs and this is just the beginning. I’m excited to see how together with Imperial we’ll shape the future of AI.”