Mr Henry Senior

Profile

Project:  Graph neural networks for scene understanding. 
Supervision: Luca Rossi  (Hong Kong Polytechnic University), Shanxin Yuan, and Prof Greg Slabaugh.

Why do a PhD?
Doing a PhD allows you to spend three years focussing on one very specific problem. If you have a question that you have been unable to satisfyingly answer during your undergraduate or masters, then a PhD is for you! It gives you the opportunity to delve into your question and work with like-minded people asking similar questions.

What are you researching and what led you work in this field of research?
My research is in the application of Graph Neural Networks to Vision-Language problems. Specifically, I work in image captioning, designing models that produce natural language descriptions of an image that go beyond simply listing the objects that appear in the scene. With over 60 million people worldwide suffering from some form of sight impairment, the development of technology to make their everyday lives easier is incredibly important. We need to develop models that can understand the visual world around us, going beyond just object recognition to models that can describe the rich interactions between things in a scene. Graphs are widely used in Computer Science and are a natural fit to representing these relationships, and therefore Graph Neural Networks are a good fit for this problem.

What is it like studying at DERI?
DERI is an amazing place to work! You’re surrounded by highly driven and bright colleagues from a vast array of backgrounds. It’s not uncommon for someone to share a paper from their field with you because it tackles a similar problem to the one you’re working on. DERI is truly interdisciplinary. The location is also beneficial, as Whitechapel has good connections to other parts of London. You also won’t struggle to find good food!

Who has influenced you along the way and whose work do you admire?
I’ve been incredibly lucky to have a lot of influential and inspiring teachers throughout my education, from my GCSE and A-Level Computing teachers to the lecturers on my undergrad and masters. During my A-Levels, I watched Stanford’s Professor Fei Fei Li’s TED talk which inspired me to think about working in Computer Vision. This, combined with my family’s history of working in optometry and orthoptics influenced my decision to work on a problem that is very impactful in terms of accessibility for those with visual impairment.

What impact do you want your research to have?
I trust that my research will help in producing models that can better understand the visual world. Current models are quite parameter inefficient, and I intend to produce model architectures that can perform well whilst being much smaller. This can be achieved by enabling them to learn about the structure within a scene.

What is the most intriguing puzzle that your research has revealed so far?
By far and away, the most intriguing puzzle in my research is “What is the best way to represent an image?”. There are lots of things to consider, from computational complexity to the types of relationships that are represented. How an image is represented can significantly impact the downstream task, which for my work means the caption might focus on certain objects rather than others. The representation of an image is an incredibly important problem for a wide range of deep learning tasks.

What online resources do you recommend?
- You and your research: https://www.youtube.com/watch?v=a1zDuOPkMSw
- PyTorch: https://pytorch.org/
- Understanding how High-Performance Computing (HPC) clusters work
- Essential reading: "How to Get a PhD, a handbook for students and their supervisors "- authors Estelle M Phillips and Colin G Johnson

Have you any key tips on keeping  motivated throughout a PhD?
Keep a regular routine. Deep rest is as important as deep work. Talk all problems (not just technical ones) through with colleagues - you'd be surprised how often they'll have had similar issues. Regular meetings with your supervisors. Pick a problem you think is important and interesting. One that you would want to work on even if you weren’t doing a PhD.

Do you have time outside your Academic work for any special interests?
Yes! I enjoy cooking and going hiking and have plenty of time for both. I’m also a member of the Lancashire Association of Bell Ringers, so whenever I’m back home I can usually be found in a bell tower ringing the changes!

Connect with Henry:
Google Scholar
LinkedIn 
GitHub 
Website

Education Background information:

• BSc(Hons) in Computer Science with Professional Experience at the University of Salford 
• MSc in Advanced Computer Science at the University of Liverpool
• Three years of industry experience as a Full Stack Software Engineer.

Research