¹û¶³Ó°Ôº

PhD in Theoretical Neuroscience and Machine Learning

The four-year programme includes in its first year intensive coursesÌýthat provide a comprehensive introduction to theoretical and systems neuroscience and machine learning (see Teaching). Multidisciplinary training in other areas of neuroscience is also available.

We offer a supportiveÌýand interdisciplinary environment with close links to the Sainsbury Wellcome Centre for Neural Circuits and Behaviours (SWC) and the ELLIS Unit at ¹û¶³Ó°Ôº. Students are encouraged to work and interact closely with researchers atÌýSWC and the ELLIS UnitÌýto benefit from this uniquely multidisciplinary environment. Projects involving collaboration with researchers at and/or external to ¹û¶³Ó°Ôº are welcome.

Students study toward a PhD in either machine learning or computational and theoretical neuroscience, with minor emphasis in the complementary field. Exceptionally, some students with pre-secured studentships have joined us to study for an MPhil degree in one of these fields. Students from other PhD programmes can also carry out all or part of their research in the unit. We do not offer taught undergraduate and Master's degree programmes, nor a research Master's degree programme.

Around 90% of our alumniÌýstudents and postdoctoral fellows work in a scientificÌýsetting, with over 60% holding an academic positionÌýand about 30% working in research developmentÌýin companies such as Google DeepMind and Facebook.

Visit the People page to learn more about our faculty and PhD students and the Research page for a brief overview of some common research themes in the unit.

Programme structure

Year 1

Our PhD programme starts with a year of intensive courses in research and techniques in neuroscience and machine learning. New students are admitted at the start of each academic year in September.

All students begin by attending Boot Camp, a three-week course covering the fundamentals of neuroscience and technology. Students will then take core courses in systems neuroscience (taught with SWC) and machine learning (incl. probabilistic machine learning and approximate inference as well as Kernels) in term 1 and theoretical neuroscience (TN) in term 2. See Teaching for more information on the courses. Students are expected to achieve a passing mark on both coursework and exams to progress to the next phase of the programme. After completing the courses, students will start developing their thesis project under the guidance of their supervisor.

Year 2

Students will work on their thesis project. They will also need to carry out a 2-month rotation project in a field different from the field of their PhD thesis (for example, if the thesis project is on TN, the rotation project should be on experimental neuroscience or ML). At the end of Year 2, students are expected to start writing their MPhil/PhD upgrade report and schedule to give an upgrade presentation to the whole unit.

Year 3 and 4

Having passed all the upgrade assessments, students will transfer from MPhil to PhD status and devote their remaining tenure to research. (Occasionally, some students may opt to only do an MPhil.)

At the end of Year 4, students will transfer to Completing Research Status (CRS). Thesis should be submitted no later than six months into the CRS period.

Across years

We foster a highly collaborative and interdisciplinary research environment. Although each student works primarily with one supervisor, we strive to avoid boundaries between groups and promote an open research culture where regular unit-wide events allow everyone to get involved in everything and be aware of what others are doing. Students thus have ready access to all members of the unit. Our external seminar series provides an opportunity to meet eminent speakers and external attendees. During their PhD, students are expected to attend all unit-wide events, including research talks, tea talks, one of the journal clubs and external seminars.

Students are encouraged to network with peers at SWC, CSML and other ¹û¶³Ó°Ôº departments (such as Institute of Cognitive Neuroscience, Institute of Neurology, Experimental Psychology; Neuroscience, Physiology and Pharmacology; Statistical Science) and attend relevant conferences such as COSYNE, ICML, NeurIPS and ICLR.

We are committed to support the career and professional development of our students. Students develop communication skills by writing scientific publications and giving regular internal talks and presentations in conferences. Critical thinking is further developed through weekly journal clubs and lab meetings. Students are expected to teach during the second year of the programme and act as a mentor to junior PhDs. In addition, students are encouraged to take advantage of the wide range of research and transferable skills courses available through the ¹û¶³Ó°Ôº Doctoral School and engage in outreach and public engagement activities (often jointly with SWC).

Entry reqruirements

Applicants must have a strong analytical and mathematical background, a keen interest in neuroscience and/or machine learning, and a relevant first degree (for example, in Computer Science, Engineering, Physics, Statistics, Mathematics, Neuroscience, or Cognitive Psychology). Students seeking to combine work in neuroscience and machine learning are particularly encouraged to apply.

Candidates offered a place on the Gatsby Unit PhD programme will be required to meet ¹û¶³Ó°Ôº's standard admissions requirements. This includes ¹û¶³Ó°Ôº's English language proficiency requirements; the required English language level for this programme is Level 3. Please review the relevant ¹û¶³Ó°Ôº web pages carefully before applying.

Studentships

Full funding is available to all students, regardless of nationality. Our PhD studentships cover ¹û¶³Ó°Ôº tuition fees for both home and international students and include an annual tax-free stipend as well as travel budget for attending conferences and workshops. We also welcomeÌýapplications from students with pre-secured funding or who are currently applying for other scholarships/studentships.

Key differences between the Gatsby Unit programme and the GCNU-SWC joint PhD programme

You can only apply to one programme (out of the three offered by the Gatsby Unit and SWC) at a given year.

Ìý	Gatsby Unit PhD programme	GCNU-SWC Joint PhD programme
First-year core courses	Systems & Theoretical Neuroscience (SNTN) Probabilistic and Unsupervised Learning; Approximate Inference and Learning in Probabilistic Models (PUL/AIL) Kernels Methods Theoretical Neuroscience (TN)	SNTN PUL/AIL Experimental Neuroscience (taught by SWC; to attend lectures & demonstrations only) TN
Rotation	Students will complete one rotation in a field different from the field of the PhD thesis i.e. if the thesis work is in TN, the rotation will be in experimental neuroscience or ML.	Students will complete two rotations, each co-supervised by one SWC experimentalist and one Gatsby Unit theorist. One pair of rotation supervisors will likely go on to co-supervise the PhD.
Supervision	Each student has a Supervisory Panel consisting of Principal and Subsidiary Supervisors. Students work primarily with their Principal Supervisors; regular unit-wide research activities ensure students have ready access to all faculty members.	Each student is co-supervised by one SWC faculty member and one Gatsby Unit faculty member. The project will involve modelling and analysing data from the lab of their SWC supervisor, in close collaboration with the Gatsby Unit supervisor.
Webpage	Ìý	See GCNU-SWC joint PhD programme

If you have additional questions, please see our FAQs.