Unleashing the potential of AI via multi-concept self-supervised representation learning

Recent advances in Self-Supervised Learning (SSL) have shown promising directions in learning robust representations without requiring labeled data. However, a nuanced challenge emerges when downstream tasks necessitate distinct levels of feature granularity. 

This PhD project seeks an ambitious candidate to develop cutting-edge SSL algorithms that can inherently capture and disentangle features at multiple granularities from image data. The research will focus on the creation of novel methodologies that not only generalise across broad categories but also discern subtle distinctions within classes, catering to a diverse array of applications. 

You will be joining the Centre for Vision Speech and Signal Processing (CVSSP) number one in the UK for computer vision research. 

Open to both UK and international candidates.

Up to 30% of our UKRI-funded studentships can be awarded to candidates paying international rate fees. Find out more about eligibility.

You will need to meet the minimum entry requirements for our PhD programme.

​​Applicants should have:  

• ​A first class (or equivalent overseas qualification) in an appropriate discipline (e.g., engineering, computer science, signal processing, physics or applied mathematics) or MSc with Distinction (or above 70% average)  
• ​You should be able to demonstrate excellent mathematical, analytic, programming skills.  
• ​Previous experience in computer vision, machine/deep learning, would be advantageous.