This page covers some frequently asked questions about applying for PhD students within the VAS research group.
Q: Are you taking PhD Students in 2023?
A: Yes, we plan to accept 2 PhD students beginning in October 2023, although this depends on funding.
Q: What topics are you working on?
A: The group works on topics on Safe Artificial Intelligence. The overarching aims of our efforts is to develop methods and tools to make AI more reliable and safer for society to use. To this end we work both on machine learning systems and logic-based systems. Topics of ongoing research include (but are not limited to): verification of neural networks and neural systems, including perception systems, verification of swarm systems including parameterised model checking, safe reinforcement learning, robust training for neural classifiers. For example, see some recent papers below on these topics (more papers are available on our publications page):
P. Henriksen, A. Lomuscio. DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis. Proceedings of the 29th International Joint Conference on Artificial Intelligence. Virtual. 2021.
P. Kouvaros, A. Lomuscio. Towards Scalable Complete Verification of Relu Neural Networks via Dependency-based Branching. Proceedings of the 29th International Joint Conference on Artificial Intelligence. Virtual. 2021.
J. Lan, Y. Zheng, A. Lomuscio. Tight Neural Network Verification via Semidefinite Relaxations and Linear Reformulations. Proceedings of the 36th AAAI Conference on Artificial Intelligence (AAAI22). Virtual. 2022
Q: What are the funding opportunities for a PhD?
Admission to PhD studies and funding are separate processes. We normally discuss funding only after a successful application for PhD studies.
Funding is competitive and it is avaliable both to selected UK and overseas students. For overseas applicants, please refer to the Imperial webpage on postgraduate study for information about tuition fees and living costs in the UK. A full list of scholarships and studentships is available here, including a new scheme from The China Scholarship Council for Chinese students at Imperial and Imperial students visiting China. Additional funding opportunities are listed here.
Admission to PhD studies and funding for PhD are treated separately.
Find a PhD provides generic information on PhD funding.
Q: How can I apply?
A: To pursue a PhD in the VAS group, you can apply directly to the Department of Computing of Imperial: https://www.imperial.ac.uk/computing/prospective-students/phd/
stating an interest in the topics we pursue.
We also welcome PhD applications on a number of available projects available on the UKRI Centre for Doctoral Training in Safe and Trusted Artificial Intelligence, which can be found here.
Q: What is the selection process?
A: Applicants are expected to have a First Class or Distinction Masters level degree, or equivalent, in a relevant, math-intensive discipline, such as computer science, mathematics, or physics. In your application you will also need to demonstrate that you have met the College’s English-language requirement, and to submit a CV and research statement. Advice about submitting an application can be found here.
Your application can only be considered after you have formally applied and after the eligibility criteria has been checked by the relevant office.
If your application is selected for assessment by the VAS group, we normally follow the following steps with all applicants:
A technical assessment based on a number of papers relevant to the PhD application. Applicants will receive papers relevant to the topic identified and asked to answer a number of technical questions on them.
Technical presentations on 2 or 3 papers related to the applicant’s research interests.
An interview on your research interests and plans.
Q: Can you help me with my research statement for the PhD application?
A: The PhD proposal is part of the application and we cannot help you with its drafting. An informative research proposal typically consists of a problem statement, followed by a literature review, an evaluation of its limitations, and a sketch of a direction of work. Note that a PhD proposal is not a binding document; it is often the case that students, in consultation with their advisor, alter the direction of their work after they begin their studies.
Q: I would like to research on X as part of your group. Do you see that as being relevant to your research agenda?
We are generally interested in Safe AI in the broadest sense. To this end we have expertise and actively work both in logic-based methods and on techniques targetting ML systems. Our present research efforts are listed above. We are open to other topics, related to the topics above. If you are still unsure, we would recommend you first submit your full application and if it is routed to us we can discuss this further.
Q: What is the difference between a standard PhD position in the group and a PhD position associated with the Safe and Trusted AI (STAI) CDT?
A: The format and progression of a student’s personal research project is the same for both standard and CDT students. There are a few differences.
CDT funding is for 4 years; Imperial funding for PhD studies is normally 3.5 years;
The STAI CDT has a number of group activities, including taught modules, group projects, hacktowns, journal club. Some but not all of these may be present in a standard PhD. For more information on CDT training see https://safeandtrustedai.org/programme/.
Note that admission to the CDT requires successful applications to both the CDT itself, and Imperial College London (or King’s College London for projects based there); details of this process are laid out here. Note that a full application on Imperial’s web site, including a research statement, is required even if applying via the STAI CDT.
Q: How often do PhD students meet their supervisor?
A: Initially students meet weekly for 30mins to 60mins to discuss a direction of work. When this is established, shorter weekly meetings are alternated with longer technical meetings as required. The group meets twice a week to discuss research work, novel papers, etc.
Q: How is Brexit affecting studying at Imperial College London?
A: The College has dedicated pages which are regularly updated with the latest information here.
Q: That’s great - should I email you before I apply?
A: If you have any specific questions not covered here, please feel free to get in touch with Alessio Lomuscio. Otherwise, we recommend you submit a full PhD application for full consideration.