Machine Learning and Artificial Intelligence (AI): Methodological Advances in Machine Learing, Generative AI and Reinforcement Learning

Machine Learning and Artificial Intelligence (AI): Flexible research project extending state of the art methodology in any area of Machine Learning, Artificial Intelligence (AI) and Reinforcement Learning.

Please contact Amos Storkey ([Email Address Removed]) as soon as possible about this opportunity. Do not wait for the deadline. See below for required information.

A fully funded Machine Learning and AI PhD position to work with Prof Amos Storkey in the School of Informatics at the University of Edinburgh, on a project titled “Methodological Advances in Machine Learning, AI or Reinforcement Learning”. The student will be an integral part of the Bayesian and Neural Systems Group. The research will forward Machine Learning Methods in any area relevant to or potential relevant to the groups interests.

We are looking for a UK national, UK resident or someone with pre/settled status (as limited by the funding).

The student should have an excellent background in Mathematics, Physics, Engineering or Computer Science, Economics, Machine Learning or Artificial Intelligence (AI), should have a significant prior training in Machine Learning methodological development. In particular they should have knowledge of: neural networks and stochastic gradient methods, Gaussian processes and kernel methods, Bayesian probabilistic approaches (inc. variational approaches and MCMC). Ideally the student would have a deeper understanding of approaches in this area.

In this role, you’ll be part of an exciting journey to advance Machine Learning and AI methodology and state of the art, particularly in methods like Deep Learning, Bayesian Modelling, and Reinforcement Learning, while re-envisioning how these technologies are developed and deployed. One particular interest is in distributed autonomous machine learning methods, but the project is flexible regarding topic. Students should have applicable project experience. Written first author publications are helpful in demonstrating this, but are not required. Good mathematical understanding and good demonstrable programming skills, ideally in Machine Learning and Artificial Intelligence (AI) Frameworks (e.g. in PyTorch) are a requirement. The project could be more theoretical or more methodological, but either way there is a strong emphasis on real world relevance of AI methods.

The awarded PhD student will be a student in the Bayesian and Neural Systems Group in the School of Informatics, University of Edinburgh. Informatics has been at the forefront on Machine Learning and AI research, being a leading institution in the field for decades (we have just celebrated a 60 year anniversary of Artificial Intelligence (AI) research at the University of Edinburgh). Recent research in the group can be seen at https://www.bayeswatch.com/publications/ .

Machine Learning and Artificial Intelligence (AI) research is still progressing at a fast pace, and so the student will need to be dedicated to working in a fast paced setting, collaborating with other Machine Learning and Artificial Intelligence students, both within the group and with the wider set of researchers in the School of Informatics and beyond. We expect students to publish in top venues in the AI field, including Neurips, ICML, ICLR and AISTATS. Interaction with other AI groups, AI-related workshops etc. will also be important.

Though this is a methodological Machine Learning and Artificial Intelligence (AI) project, the work will need to be practically demonstrated. The potential application areas vary across the whole range of Artificial Intelligence (AI) applications, including computer vision to natural language processing. The student can propose a potential direction for the research and we will use that as a springboard to finalise a research project.

As a group we spend time pioneering new Machine Learning, AI and deep learning methods to make modern Machine Learning models, foundation models and generative AI techniques more efficient, more effective and more distributed. We have an ambition to make Artificial Intelligence methods (AI) rest in the hands of the people rather than in the hands of large companies.

Candidate’s profile

• a strong degree or higher qualification in a relevant field (e.g. artificial intelligence (AI), machine learning, computer science, maths/mathematics, engineering, physical sciences, economics or any other field where evidence is provided of sufficient computing and mathematical background)
• solid experience of programming, learning methods and ideally AI deep learning environments (e.g. pytorch) or a computer systems background
• good mathematical skills
• demonstrable writing capability.

Studentship and eligibility

This funded post is suitable for a home student (e.g. students ordinarily resident in Scotland or the rest of the UK – England, Wales or Northern Ireland, Republic of Ireland, and EU-EEA nationals with Pre/Settled status). We are particularly supportive of applications from members of communities that are under-represented in machine learning.

Application information

We advise eligible and potentially interested students to contact Amos Storkey ([Email Address Removed]), Professor of Machine Learning and Artificial Intelligence (AI) as soon as possible with a CV and statement of research interest for more information and transcripts if you have them.

Applicants may choose to fill in the form at https://forms.office.com/e/QpuYH8mwEw to help demonstrate their proficient deeper understanding of machine learning and AI. This is not required, but applicants may find it interesting.

Environment

The School of Informatics is one of the largest in Europe and currently the top Informatics institute in the UK for research power, with 40% of its research outputs considered world-leading (top grade), and almost 50% considered top grade for societal impact. The University of Edinburgh is constantly ranked among the world’s top universities and is a highly international environment with several centres of excellence. It has been researching Artificial Intelligence (AI) for over 60 years, and has pioneered many key machine learning and AI methods. It has great strengths in computer vision and natural language processing.

Example project areas: AI and Bayesian Methods, Robust AI Capability, AI Trustworthiness, Solving Continual Learning, Efficient AI Training, Adversarial approaches in AI, accurate uncertainty quantification in AI methods, multiagent AI, Machine Learning Markets, Transferable AI models, Enabling AI to cope with domain shift, AI foundation models, AI Architecture Search, flexible AI architectures, real-time AI. AI personalisation and deployment, interactive AI, integrating different AI models, continuously adaptive AI methods, AI and long horizon planning, theoretical foundations for AI, value of information in AI models, optimal model reuse etc.