Distributed Machine Learning over Mobile Networks PhD

This is a full-time PhD opportunity sponsored by Cranfield University under the UK's EPSRC Future Telecommunications Hubs (FTH) ecosystem, which provides training and research as part of a larger ecosystem. The financial offering includes UK home fees and a bursary. The focus area is distributed AI over radio access networks.

This project focuses on how AI will provide a seamless service to mobile users of tomorrow. The student is expected to have a background in AI with a curiosity about wireless networking.

Machine learning that occurs over a distributed network is attractive to reduce latency, ensure data and compute is sovereignly hosted, and utilise available edge computation resources. There are many challenges when the demand is spatially and temporally mobile, and when the inference is computationally heavy. This PhD studentship will advance research in AI on Radio Access Networks (AI-on-RAN) by exploring how to optimise distributed and mobile inference of large language models by adapting how tokens, memory, and network bandwidth is used in real time.

Cranfield University is a post-graduate only university focused on critical technologies such as aerospace, defence, AI, robotics, and telecoms. It is situated just north of London between Milton Keynes and Bedford. There is a proposed merger with King's College London in 2027.

The outcomes are expected to shape how future AI algorithms are designed, as well as how future computation and communication resources are invested in 6G telecommunication networks.

There is both university-based and wider FTH ecosystem training with other PhD students from universities across the UK. The experience gained is cross-disciplinary (AI and telecoms), with exposure to different stakeholders including defence, government, and infrastructure.

Supervisors:
1st Supervisor: Prof. Weisi Guo
2nd Supervisor: Dr. Mengbang Zou

Entry Requirements:
Applicants should have a first or second class UK honours degree or equivalent in a related discipline of computer science, statistics, or electronic/information engineering. This project would suit students with a fundamental curiosity about how foundation models work at the algorithmic level. An inclusive approach will be taken during recruitment, considering students from underrepresented groups and different socioeconomic groups.

Funding:
Sponsored as part of the EPSRC Future Telecommunications Hubs (FTH) initiative and Cranfield University, this studentship will provide a bursary of up to £21,805 (tax free) plus fees for three years. Applicants must be classified as a Home fee status student to be eligible for this funding.

Research students at Cranfield benefit from being part of the Cranfield Doctoral Network, which brings together research students and staff to share ideas, identify collaboration opportunities, and create communities of practice. Students are also expected to attend the Doctoral Researchers' Core Development (DRCD) Programme, comprising four one-week units covering key stages of the research journey.