PhD studentships at UCL in Foundational AI

[Gabriel J. Brostow]

Deadlines: 1st January 2021 in the first instance, though applications
are accepted through around 1st June 2021. Applying earlier is better!

At University College London, we have a Foundational AI CDT, which means
that every year, we accept and fund around 10 new PhD students. The
subject areas cover a spectrum of CS and related subjects, including
Vision, ML, Robotics, NLP, Graphics, and more.

The idea is that each cohort (this is the 3rd batch of ~10) works on AI
innovations for a "normal" PhD, embedded within a research group in the
department. Additionally, members of the cohort learn and work together,
to get the benefits of being in a cross-disciplinary group, selected
especially to be diverse in terms of people, specialties, and aims. You
will be expected to attend talks and workshops that prepare you
particularly for either academia or entrepreneurship.

Funding for PhD students in the UK does unfortunately take
residence/nationality into account. This program has more flexibility for
students who apply earlier. We particularly welcome female and non-binary
applicants and those from an ethnic minority, as they are
under-represented within UCL's PhD programs. We're working on that.

To apply, visit
https://www.ucl.ac.uk/ai-centre/study/cdt-foundational-ai/applying-foundational-artificial-intelligence-mphilphd.
Be sure to follow the advice there, though you may have to list PhD
supervisor names based on whom you're most aligned with, without having
met them first. The point is to get your application in front of the
supervisor(s) who align with your research passion.

---- ---- ----

Extra info 1: Other UCL PhD studentships in ML

UCL's Computer Science Department has a separate admissions process:
https://www.ucl.ac.uk/prospective-students/graduate/research-degrees/computer-science-mphil-phd
There, it's possible to be admitted and funded, or sometimes strong
candidates are admitted without funding - it depends on isolated funding
within a given research group, projects, scholarships, sponsors etc. Like
the Foundational AI CDT, or even more so, UK (or maybe EU - we shall see)
applicants are easier to fund because of national rules. That pains me
too.

Extra info 2: My research team (Gabe Brostow's)

My own research preferences lean heavily toward human-in-the-loop computer
vision and ML, where a system learns from small numbers of examples to get
better with use, and helps a person, e.g. to do their job, to learn
faster, or to adapt for different abilities. The core research must be
publishable in CVPR- or CHI- like conferences. A small fraction of
successful members from my group has gone on to startups, which is also
something I'm happy to support - but that's secondary to great research.
If I could design the perfect applicant (please don't let this intimidate
you!), they'd be a strong programmer who wants to build their experiments
with user-interfaces for people, sometimes especially for children, on top
of creative ML models, while reading, writing, and sharing exciting papers
with the rest of the team. Drive and trajectory beats a pile of
experience, but those are harder to judge. So your application needs to
please show where you're planning to go!





--
-Gabriel J. Brostow
http://www.cs.ucl.ac.uk/staff/G.Brostow