Skip to main content

Research Studentship in Efficient Uncertainty Estimation of Modern Machine Learning Systems

Research Studentship in Efficient Uncertainty Estimation of Modern Machine Learning Systems

Project: "Efficient Uncertainty Estimation of Modern Machine Learning Systems"

3.5-year DPhil studentship 

Supervisors: Prof Philip Torr

"The aim of the project is to advance calibration, out-of-distribution detection, and improve uncertainty estimates in modern machine learning systems. Deep neural networks are known to be too confident about predictions of samples that are far from the training data. Addressing this nuisance is particularly essential with the real-life deployment of such systems, e.g. self-driving cars, where a deep learning model can easily encounter situations that have not been presented during the training phase. Moreover, we seek to find efficient ways for this task. For instance, while ensembles are known to be state-of-art in calibration, they remain to suffer from the large computational overhead. This project aims to find approaches for uncertainty estimation in an efficient manner through sparsification or importance sampling over architectures among others."


This studentship is funded through the Horizon Robotics Scholarship.

Award Value

Course fees are covered, the stipend (tax-free maintenance grant) is c. £16,062 p.a. for the first year, and at least this amount for a further two and a half years. 

Candidate Requirements

Prospective candidates will be judged according to how well they meet the following criteria:

  1. A first-class honours degree in a mathematical discipline.
  2. Excellent English communication skills (oral and written).
  3. Preference will be given to those who have experience in research in machine learning or computer vision.

The following skills are desirable but not essential:

  1. Masters in machine learning or related field.
  2. Publications in top tier venues (NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, etc).
  3. Experience in programing and capacity to generate original research ideas.

Application Procedure

Informal enquiries are encouraged and should be addressed to Prof Philip Torr.

Candidates must submit a graduate application form and are expected to meet the graduate admissions criteria.  Details are available on the course page of the University website.

Please quote 22ENGIN_PT in all correspondence and in your graduate application.

Application deadline: noon on 3 June 2022

Start date: October 2022