Reflection on the British Machine Vision Association summer school

post by Muhammad Suhaib Shahid (2020 cohort)

Over the summer of 2022, I attended the annual British Machine Vision Association summer school at The University of East Anglia. The weeklong summer school provided an opportunity for computer vision researchers, in both academia and industry, to explore a wide range of computer vision topics through intensive lectures and labs. From the 11th to the 15th of July, I had the chance to learn from expert instructors, participate in hands-on activities, and connect with other professionals and students in my field. In this reflection, I will share my key takeaways and insights from the summer school, as well as any challenges or setbacks I faced and how I overcame them. I will also discuss the impact of summer school on my personal and professional development.

Over the course of the week, I had the opportunity to learn from expert researchers from some of the most active research groups in the field of computer vision both in the UK and abroad. Throughout the series of lectures, they provided comprehensive overviews of the fundamental concepts and techniques of image processing and analysis, as well as lectures on unsupervised learning, image segmentation and deep learning for Computer Vision. Though most computer vision researchers are very much adept in these areas, re-covering these subjects proved key for a complete understanding of the advanced topics that we encountered later in the program.

Over the five-day period, there were 16 lectures and one lab with each session being run by an expert lecturer or instructor in that specific field. Most relevant to my research were two afternoon sessions held on days two and three. The first headed by Oscar A. Mendez on “Deep Learning for Computer Vision” and the second by Chris Willcocks on “The concepts and characteristics of different deep generative modelling approaches”. The two sessions addressed the use of Computer Vision in medical image processing and synthesis, a topic key to my research. Deep learning, which is a type of machine learning that uses neural networks to learn from large amounts of data, was a reoccurring theme throughout the week. Many researchers, myself included, encounter issues surrounding finding large enough data sets, a problem prevalent, especially when working with medical imaging.  I found these topics to be particularly interesting, as they are at the forefront of computer vision research and have many practical applications in a variety of fields.

In addition to the lectures and discussions, we also had the opportunity to participate in a few hands-on activities that allowed us to apply what we were learning in a practical setting. The afternoon session of the first day took place in a computer lab. Working on individual tasks, through helping each other along when needed, we implemented various PyTorch functionalities from scratch. It was fascinating to see how the tools and packages used in our everyday work as researchers are implemented, something that often crosses my mind now months later whilst undertaking various programming endeavours.  I appreciated the opportunity to work with my team and receive both instruction and feedback from the instructors.

One of the most memorable experiences I had during the summer school was the opportunity to connect with other professionals and students interested in computer vision. The program brought together a diverse group of individuals from a variety of backgrounds, disciplines, and even countries. It was inspiring to see the different perspectives and approaches that people brought to the table. We had several networking events and social activities that provided us with the opportunity to get to know each other and learn about each other’s interests and goals. I made some lasting connections that I believe will be beneficial in my future academic and professional endeavours.

Overall, the summer school has had a significant impact on my personal and professional development. In addition to learning about the technical aspects of computer vision, I also gained valuable skills in networking, problem-solving, and communication. I feel more confident in my ability to design and implement computer vision systems, and I am excited to explore the many applications of this technology in the future. However, the summer school was not without its challenges and setbacks. One of the biggest challenges I faced was keeping up with the fast-paced nature of the program. There was a lot of information to absorb in a short period of time, and I found myself struggling to keep up at times. However, I was able to overcome this challenge by staying organized, asking for help when I needed it, and making the most of the resources available to me.

I do have a few suggestions for improving the summer school experience for future participants: it would be helpful to have more time for networking and connecting with professionals in the field. While I enjoyed the opportunity to meet and collaborate with other students, I also think it would be valuable to have more structured opportunities to engage with industry professionals and learn about career paths in computer vision. There was only a lecture that briefly touched on how to find the best post-research opportunities. Additionally, the time allocated to lab work was not sufficient. There was only one session, and no opportunity to implement our knowledge in any real-life scenarios; I would suggest that future events take advantage of the wide knowledge base available and allow students to work in groups, fostering greater opportunities to apply their domain-specific knowledge.

In conclusion, the summer school on computer vision was a valuable and enriching experience that has had a lasting impact on my personal and professional development. I am grateful for the opportunity to have learned from expert instructors and to have connected with other like-minded individuals. I highly recommend this summer school to anyone interested in exploring the exciting field of computer vision!

Angela meets Industry Partner to discuss their future strategy

post by Angela Thornton

I’ve just come back from 5 days in Nardo, Southern Italy where I was on retreat with my industry partner the Carboncopies Foundation (CCF). We stayed in Relais Monastero Santa Teresa – a 13th-century baronial palace which was subsequently converted into a cloistered monastery. The bedrooms were all beautifully unique and our meeting rooms were equally stunning even if getting the Wi fi to work and finding enough plug sockets and adapters could be challenging! We had a few hours off on the Friday to explore the historic town with its cobbled streets and beautiful buildings and even managed a short trip to the nearby seaside on Sunday when it was a lovely sunny day.

However, we had an intensive schedule which involved launching new departments and initiatives and also progressing existing projects. I’m on the Board and actively involved in both the main Research and Education offices as well as leading the Ethics department so there was a lot of information to discuss and disseminate. Specifically I led an ideation session on an Ethical Framework for Whole Brain Emulation (WBE) which is one of my key areas of responsibility for 2023 and beyond. This is an active research area but WBE in humans is likely to take many decades so planning to ensure responsible research and innovation is challenging. Thankfully as an ex-industry researcher, I have considerable experience in ethics and codes of conduct and was on the Ethics and Compliance Committee for the British Healthcare Business Intelligence Association. Hence I was able to draw on that experience to moderate a creative session to identify key topics and risks. Following idea generation, we conducted some interim thematic analysis which I will review and develop to provide a draft table of contents for the Ethical Framework. This first iteration is an internal process, but CCF will subsequently consult with other experts in the field with the aim of authoring the first set of Ethical Guidelines for WBE.  It was particularly productive to have members of the CCF who aren’t always directly involved in ethics as they bought a unique perspective.

The event was notable not just in how much we achieved collaboratively but how enjoyable it was to spend time with people that I had only previously met virtually. Our shared memories will include some entertaining moments including star gazing from the roof and spinning LED glow poi.

Reflection on outreach: Dstl AI Fest 4

post by Matthew Yates (2018 cohort)

In October 2021 I gave a presentation on my PhD project at Dstl’s AI Fest 4. This is a now annual event held by Dstl and attended by various government departments, industry partners and academic researchers. The event was held online over a virtual conference meetings platform with over 100 different talks from AI experts over the course of two days.

The central aim of the event was to discuss topics surrounding “Trustworthy AI” with Dstl stating their mission is “to de-mystify the area of AI by helping MOD understand how it can responsibly and ethically adopt AI in order to deter and de-escalate conflict, save lives and reduce harm.” Like other virtual conferences that become the norm during the pandemic there were different panels where attendees could interact with the speakers during Q&As as well as various networking rooms to talk to others about research.

I presented my work as part of the “AI Methods and Models” panel, with my then-current working title for my PhD and presentation being “Accurate Detection Methods for Image Synthesis”. As I had done multiple previous presentations for my PhD at various stages of the project I knew the content quite well but made some adjustments for my specific audience and what I wanted to focus on as the key message of my work. As I was presenting a work in progress, with still a year of research to go, I decided it would be best that I focus on the importance of the interdisciplinary methodologies I was using for my project rather than any final models or results (which I was still currently working on). I also thought it might help differentiate mine from the rest of the panel which was concerned with various novel implementations of machine learning models.

In my presentation, I gave a brief overview of myself and my project’s background as well as a short explanation of Generative Adversarial Networks for those in the audience who were less familiar with deep learning models. The main content of my presentation however was my mixed methods approach to looking at fake image detection. I explained how as the objective of a lot of fake image generation is to fool human visual perception (e.g. fake news, deep fakes etc) taking a human-centric approach to investigating novel detection methods is equally, if not more important, than looking at purely algorithmic solutions. I then presented the results of my initial image detection study which found differences in detection behaviour between computational and human detection methods as well as differences between experts and novices. As this was the part of my research plan I was currently up to, I spent the rest of the presentation discussing the implications for my current results and how they were going to inform the rest of my work. At the end of the presentation, I took a short Q&A with questions about possible other metrics I could use to measure human visual perception towards image detection like eye tracking and also how to combine this research with automated methods. Both of these questions were easy for me to answer as fortunately they were both ideas that I had planned to explore myself at the final stage of my research.

On reflection, I thought I had gotten my main points across in an engaging way and had been able to communicate it to people of differing levels of technical backgrounds, however, with it being held online it was much more difficult to get a sense of this than if I had been presenting face to face. Although presenting to a screen can sometimes alleviate any nerves in presenting to a large, live audience, I find it can also be quite hard when you don’t have any visual feedback from audience members present.

Despite some of the reservations I still have towards online conferences I did find my experience presenting at Dstl’s AI Fest useful. In addition to the experience of having to communicate my research to a live audience, it was also a useful opportunity to get to know other people either at Dstl or in industry. The timing of the event coincided with my 3-month Dstl internship so some of the people I was working with at the time also attended the conference and could get an idea of what kind of work I was doing at Horizon and the Computer Vision Lab for my PhD project.

 

Presenting the Future of Healthcare at the Cobot Maker Space

post by Angela Higgins (2022 cohort)

Students from the Horizon Centre for Doctoral Training were hosted by the Cobot Maker Space to present their visions of healthcare utopias, and how to avoid ending up in a medical dystopia.

From robotics and AI to health tracking and gene sequencing, are we headed towards a utopian or dystopian future for healthcare? For Future Products Sector Day at the Horizon CDT programme, our group were asked to present our future visions of healthcare to the PhD cohort. Using the range of robots available in the Cobot space, we demonstrated and discussed how these technologies could be used to the benefit or detriment of human wellbeing in the future.

Demonstrations included a UV cleaning robot which could be used to disinfect hospitals, telepresence robots which could be used for remote doctors’ visits, and companion robots for older people. In a walkaround tour of the living space, an experimental area designed to simulate a home living room and kitchen, we demonstrated how Internet of Things sensors could be used to monitor activity and help people track their health. These technologies have great potential to allow people to take control of their wellbeing, keep medical professionals in the loop and ultimately allow people to live in their own homes for longer. However, this raises questions about surveillance, data protection, privacy, and dignity for older people.

Academics from the University of Nottingham contributed their expertise, including Professor Praminda Caleb-Solly, who spoke about her work researching robots to help support older people. Praminda spoke about work at the CHART research group, and how robots could be used to enable and enhance human-human interaction for health and care, rather than replace it. Epidemiology PhD student Salma Almidani also spoke via video interview, discussing future pandemics, vaccine hesitancy, and where technologies may be useful in healthcare ecosystems of the future.

The afternoon was finished off with the talks from us, in the 2022 cohort. Jon Chaloner spoke about universal health coverage and how we can work towards providing everyone worldwide with a full range of accessible healthcare across their lifetime. Gift Odoh talked about how healthcare and telepresence robotics can influence and benefit each other, through technology and knowledge exchange. Finally, I closed by talking about the future of pain, and how we could use robotic devices and sensing technology to better understand and respond to and manage pain.

The afternoon provided a range of emerging perspectives on the future of healthcare, and debate about how these technologies could be used and abused. Ultimately, by discussing and exploring imagined utopias of the future, perhaps we can identify routes to get there, whilst avoiding some of the dystopic pitfalls along the way.

‘Interactions with Coffee Wizard’: Reflections on a first conference paper

post by Oliver Miles (2018 cohort)

Overview

In my first paper, ‘Interactions with CoffeeWizard’ I gave an account of participant interaction with a values-orientated choice and prediction framework, embedded through a coffee selection box activity in the home. It was originally submitted for the Computer Supported Cooperative Work (CSCW) 2021 conference, but despite feedback citing key points of merit including good alignment venue, a sound methodology, and agreement with the general line of argument, the paper was unfortunately rejected at first pass. Despite reading and re-reading the reviewer’s comments at the time, it is only now during the process of writing up a new study based on the same framework that I fully appreciate and can apply the suggested changes.

In the following, I will give an overview of the paper, the motivation for writing it and for specifically choosing the CSCW venue. I will then reflect on the more practical points of collaboration and advice during paper drafting, before outlining some of the feedback I received and how I intend to improve my next submission based on this. Hopefully, sharing some of my insights regarding responding to reviews can help others – particularly if they are writing their first paper as a solo author.

‘Interactions with CoffeeWizard’

Broadly, my thesis explores the use of discrete value sets such as product attributes and personal end-goals in life, as grounds for personalization in the recommendation of everyday coffee consumption. The purpose of my first study was to deploy a novel interaction framework for surveying, predicting, and eliciting retrospection on personal value preferences, elicited through coffee choice selection and reflected the participant as infographics. Realised through the initial questionnaire, coffee selection activity, and follow-up interview, this would allow me to demonstrate the kinds of interaction elicited at each stage of the framework and improve the proposition so that it captures this as rich, contextual data. The study was delivered as a domestic deployment due to Covid-19 restrictions at the time. In the findings, I present and discuss the results of interviews with 12 participants, whose reflections enabled a discussion based on the emergent, practical values of selecting coffee based on its reputed value attributes and making choices incongruent/congruent with apparent predictions.

Motivations

In terms of motivation for the paper, I wanted to share my theoretical ideas with peers in the human computer interaction (HCI) community whose work tends towards testing and developing prototypes. I chose the CSCW conference as it positions itself as ‘…a premier venue for presenting research in the design and use of technologies that affect groups, organizations, and communities’[1], which is well aligned with the practicalities of incorporating social research and HCI methodology. More broadly, this was my first opportunity to formally share ideas with an HCI audience whose common idioms can be challenging to adopt when approaching the field from another discipline.

Paper preparation

If you are completely new to paper writing like I was for this piece, I would recommend attending any of the Research Academy courses related to effective writing as soon as you get the opportunity. Paper writing is fundamentally different in my experience from other forms of academic writing as it requires your work to retain its originality while at the same time reflecting the nuances of the conference or journal, not to mention strict formatting and editorial guidelines.  Supervisor feedback is therefore also crucial during the drafting process.  It can be tempting to wait until you have entire finished sections or even a full draft before you seek feedback. To counter this, I have found that the following ‘skeleton paper’ approach works for me:

    • Outline the known section headings from ‘introduction’ to ‘conclusion’
    • Break these down further into the main substantive points you wish to cover
    • Ensure there is a clear narrative that will bring your reader to the intended contribution

This can effectively read as a full draft while allowing efficient iterations, which can be further substantiated once the main concepts and contributions become coherent.

Handling Reviewers’ Comments

It is easy in hindsight to see the extent to which I was/was not following my own advice regarding paper preparation. On the one hand, reviewers picked up on some key merits which resonated with my intended contribution: The proposition appeared to give ‘insightful’ findings; the issue of value-based personalization was agreed as a relevant interactional one, and the methodology itself was judged to be appropriate. These points were generally very encouraging given the wider implication of the validity of my thesis.

Nevertheless, under-developed contributions and literature selections were significant enough to result in rejection. In the first case, I reflect that I had not invested enough in the preparation of the paper specifically, a paper for the CSCW audience. This left reviewers with a sense that they were being left to draw out findings relevant to them, instead of having them clearly outlined. This is intrinsically linked to the second concern regarding literature selection. I had only referenced one CSCW publication, with the rest of my sources coming from other conferences or journals. So, while the material I based my work on was described as ‘relevant’, it made it difficult for reviewers to link any contribution back to work specifically emanating from the venue itself.

Role of paper within PhD

‘Interactions with CoffeeWizard’ continues to play a significant role in my thesis as the first deployment of a novel, values-orientated personalization framework. Expanding on the literature section and aligning the contribution more closely with contemporary works from CSCW, the work now forms the first empirical chapter of my thesis. In this sense, I hope that reviewer feedback has improved the communication of my work for my PhD itself, as well as informed me how I am currently approaching write-up of my final study to a similar venue in 2023.

[1] https://dl.acm.org/conference/cscw

Call for Participants: STUDY OF THE PSYCHOLOGY OF EXCESSIVE VIEWING ON DEMAND

post by Joanne Parkes (2020 cohort)

As some of you know, I’m a 3rd-year Horizon CDT PhD student partnered with BBC Research & Development and based within N/Lab at the University of Nottingham. I am researching binge-watching behaviours and how we might better manage them if they’re problematic.

Purpose: For this study, I am looking for participants to take part in a 1:1 interview via an online Teams meeting to discuss their viewing habits, perspectives on binge watching and thoughts on why/when people might watch more than they intend.

Who can participate? This study is open to anyone aged 18 and over who regularly (typically at least once a week) watches 2 or more episodes of the same programme and/or 2 or more continuous hours of on-demand television as their main activity.

Commitment: The interview should take around 60 minutes to complete.

Reward: £15.00 Amazon e-voucher for your participation.

How to participate: Email me at joanne.parkes@nottingham.ac.uk to express your interest and arrange a mutually convenient meeting time. Evenings and weekends will also be available.

More information is available. For any queries, please feel free to contact me using the email address provided.

Read more about my research project.

Call for Participants: Fake Image Detection w/ Eye tracking

post by Matthew Yates (2018 cohort)

I am a final year Horizon CDT PhD student partnered with the Dstl. My PhD project is about the detection of deep learning-generated aerial images, with the final goal of improving current detection models.

For my study, I am looking for participants to take part in my short face-to-face study on detecting fake aerial images. We have used Generative Adversarial Networks (GANs) to create these.

I am looking for participants from all backgrounds, as well as those who have specific experience in dealing with either Earth Observation Data (e.g. aerial imagery, satellite images) or GAN-generated images.

Purpose: To capture gaze behaviour during the detection of GAN generated fake images from real aerial photos of rural and urban environments. Participant accuracy and eye movements will be recorded.

Who can participate? This is open to anyone who would like to take part, although the involvement of people with experience dealing with related image data (e.g. satellite images, GAN images) is of particular interest.

Commitment: The study should take between 20-40 mins to complete and takes place in Room B73 Computer Science Building Jubilee Campus

Reward: £10 amazon voucher for your participation

How to participate? Email me at matthew.yates1@nottingham.ac.uk with dates/times that you are free to arrange a timeslot

For any additional information or queries please feel free to contact me,

Thanks for your time,

Matt

+44 (0) 747 386 1599 matthew.yates1@nottingham.ac.uk 

Map Check

post by Vincent Bryce (2019 cohort)

This summer saw me walking St Cuthbert’s Way, a 100km hiking trail in the Scottish Borders/Northumbria area with my children. It was a great trip, plenty of challenge but achievable, and I’d recommend it:  https://www.stcuthbertsway.info/long-distance-route/

The trail was well signed, but needed us to use our map and compass in places. It’s three years since I started the PhD, two of which were part-time, and it feels like a good time to check where I am and where I’m going:

Where am I

I’m starting the third year of a part-time PhD in the Horizon CDT, focussing on responsible research and innovation (RRI) and Human Resource information systems. This is about exploring how organisations can innovate responsibly with digital technologies, the challenges this involves, and some of the specific issues for HR technologies.

I’ve chosen a Thesis by concurrent publications route => a set of related studies rather than one overarching thesis.

Where am I going

I am going to complete my PhD and plan to come back into full time HR work, applying the insights into my digital HR work. The experience of being a student and researcher at the University I work will help me keep a strong customer focus.

What have I done so far

Following a year of taught activity about a range of digital economy and computer science topics, I’ve completed a series of studies and articles.

Highlights include a study on published Responsible Innovation case studies exploring the benefits of RRI, pieces on HR analytics and their ethical implications, presenting at the Ethicomp, CIPD Applied Research, and Philosophy of Management conferences, and critical articles on wider challenges for responsible innovation such as low-code technologies and crosscultural aspects.

I’ve seen new ideas and emerging technologies, and built skills in cding, data science, writingm writing to data science, and from bot based blogging, digital watercoolers and AI coaching to augmented and virtual reality tools.

What are my main findings to date

  • Responsible innovation practices are associated with business benefits.
  • Digital technologies, in particular ones users can reconfigure for themselves, pose challenges for responsible innovation methodologies, because these tend to rely on the technology being developed in ways which anticipate and respond to societal needs. End users, rather than scientists and developers, are increasingly able to innovate for themselves.
  • Algorithmic HR technologies give HR new capabilities, but are linked to some ethical concerns and have features which imply a need for responsible innovation and implementation.
  • Interviews with HRIS suppliers have limited opportunities to engage wider stakeholders and anticipate downstream impacts, creating reliance on client organisations to reflect on how they apply the technologies.
  • The knowledge and values of HR practitioners are a critical constraint on responsible algorithmic HR adoption.

What are my priorities for the coming year

Completing my thesis synthesis document; concluding in-progress studies on the increasing scope of employee data collection, and HRIS supplier and practitioner perspectives; and getting in position to submit by Sep 2023.

Right – onwards! I’ve recently attended the Productivity & the Futures of Work GRP conference on Artificial Intelligence and Digital Technologies in the Workplace to present about my study on the increasing scope of employee data collection and hear about what’s hot and what’s not.

originally posted on Vincent’s blog

What is Pint of Science?

post by Peter Boyes (2018 cohort)

“The Pint of Science festival aims to deliver interesting and relevant talks on the latest science research in an accessible format to the public – mainly across bars, pubs, cafes and other public spaces. We want to provide a platform which allows people to discuss research with the people who carry it out and no prior knowledge of the subject is required.”

This was a new one for me, a collision of worlds. I’ve spent the last 8 years in Nottingham, studying for my undergraduate, master’s, and now PhD. I did some extra bits with my course, PASS leader in its first year in the school of mathematics, and some events here and there as a PhD researcher, but I stick mostly to my studies and then explore volunteering in places beyond academia. I’ve enjoyed helping coordinate sports clubs and competitions since joining university, but Pint of Science arose as an opportunity to combine my two halves. Volunteering and putting on events related to my studying.

I got involved at first as a general body to lend a hand on a couple of the nights, but moved into a Theme Lead role early on in the year when an opening popped up. About 9 months ago myself and my team of fellow volunteers were allocated our theme (Beautiful Mind – anything around human senses and the brain) and we set about recruiting speakers and planning our event. We had 3 evenings at Bunkers Hill, Hockley to fill, and grouped our 9 speakers into similar topic areas. These topics covered broadly Pain, Senses, and Mental Health. We checked out the venue space, and planned out schedules for the nights, with presentations, Q&As, and some activities for the audience such as quizzes (what else do you expect on a weeknight in a pub when you’re talking science). May flew round, and tickets got snapped up. The nights went fantastically, there was a buzz with the great speakers and the final night in particular packed out the venue space to end on a high note.

This side venture was a little outside my comfort zone, yes I’m familiar with volunteering and running events, and I’ve been in academia for 8 years, but the theme wasn’t in my area of expertise and science outreach is a new experience for me. I was supported well throughout by a great team more familiar with the topics and events like this one. I’ve learned a lot about outreach through these nights. This was for me about learning how to facilitate public outreach and conveying cutting edge research and expert topics to the general public, no easy task. The most revealing part of each night was being able to listen to the speakers talking to each other, some seasoned Pint of Science-ers, some new to the event. I also had the privilege of facilitating fellow Horizon CDT 2018 cohort member Shazmin Majid presenting her latest work.

This experience has given me confidence in presenting my work and how to go about it, equally how not to go about it. Avoid overloading slides with text and too much inaccessible specialist terminology. It’s fine to use some if you define them and get the audience up to speed, but need to find other ways to convey your research if every slide needs 5 terms defining or sub definitions, it breaks up any flow and makes it difficult to follow, particularly for non-experts in the field. Analogies are great, again not too many, and not too convoluted. I have been given advice before on using analogies as they can lead to misunderstanding of concepts if followed too afar, but a well-crafted one can enhance the audience understanding. Demonstrations or activities that let the audience learn through involvement rather than relying on a perfect explanation also seal the deal on a great outreach talk. The simpler the demo the more effective. Though, doing any of those things is no easy task.

I would encourage other CDT students to get involved in the coming years from either side, later stage PhD students and recently graduated alumni have a great opportunity to put your work out there. Early stage candidates should see how other researchers slightly further along your journey are engaging with this sort of outreach, it might even give you ideas about your own research.

 

 

 

Summer School Participation Reflection

post by Matthew Yates (2018 cohort)

I participated in the 2022 BMVA Computer Vision summer school which was held at The University of East Anglia. The summer school was aimed at PhD students and early-stage researchers who were involved in the research areas of computer vision, image processing, data science and machine learning. The event was held between the 11th – 15th of July and consisted of a full week of lectures from a variety of researchers who are experts in the field on a wide array of computer vision topics with an emphasis on the latest trends that are pushing the field forward. In addition to the lectures, there was also a programming workshop and social activities in the evenings such as a computer vision-themed pub quiz and a dinner held at Norwich Cathedral.

As the lectures covered a wide range of topics not all of them were strictly relevant to my own PhD project and research interests, although it was useful to be exposed to these other areas to gain some tangential knowledge. The event started with a lecture by Ulrik Beierholm on cognitive vision, how it functions and how it compares and contrasts with similar computational vision systems such as Convolutional Neural Networks (CNNs). As my own background is in cognitive psychology and computational neuroscience, I found the lecture very engaging, even if it was mainly reiterating ideas I had already studied during my Masters’ degree. The afternoon of the first day was given to a programming workshop where we were given tasks on a Google Colab document to help familiarise ourselves with using PyTorch and also programming some of the key parts of a deep learning model pipeline. Although these were fun and useful tasks, we were not given enough time to complete them as much of the first half of the workshop was taken up with technical issues in setting up the guest accounts to use the labs computers.

The second day started and finished later than the first, with more lectures and an event in the evening, a structure followed throughout the rest of the summer school. The first lecture of the day was on Colour by Maria Vanrell Martoreli. Going into this lecture with no expectations I came out of it having found it very useful, with a much deeper understanding of the role of colour in the interpretation of objects in an image, both in human and machine vision systems. These were followed by lectures on image segmentation by Xianghua Xie and local feature descriptors by Krystian Mikolajczyk.

The image segmentation lecture presented some of the latest methods being used as well as some of the common problems and pitfalls encountered by researchers implementing these methods. While these two lectures presented a lot of well-articulated ideas in their respective areas, they were fell out of my own research interests so I don’t think I got as much value out of them as others in the room.

The last lecture of the day was a rather densely packed overview of deep learning for computer vision by Oscar A Mendez. This was a very engaging lecture and with a lot of information including some good refreshers on more fundamental architectures such as MLPs and CNNs and a very intuitive introduction to Transformers, a rather complex deep learning model which are currently very popular in many research areas. In the evening we went into Norwich city centre for a bowling social event.

Wednesday morning consisted of lectures on Shape and Appearance models by Tim Cootes and Uncertainty in Vision by Neill Campbell. Both of these were conducted online over teams due to the presenters having caught covid after attending the CVPR conference the previous week. The shape and appearance models lecture was informative but not of much particular interest to me but the uncertainty in vision was quite interesting and the presenter managed to include a good level of audience engagement activities despite being over a webcam.

After lunch we had a lecture on generative modelling by Chris Willcocks. This was a very interesting lecture as it covered the current trends in generative modelling (e.g., GANs, Transformers) and also looking at the architectures which have the potential to be the future of the field such as diffusion and implicit networks. As my own work looks at GANs, I found this talk to be particularly enlightening and also comforting as it agreed with many of my own arguments, I include in my thesis such as the current issues with using FID as an evaluation metric. In the evening we attended a dinner at Norwich Cathedral which gave everyone a good time to network and discuss the week’s events with other members of the summer school.

Thursday consisted of another full day of lectures on various topics in Computer Vision. These were Unsupervised learning by Christian Rupprecht, Structured generative models for vision by Paul Henderson, 4D performance capture by Armin Mustafa, Egocentric vision by Michael Wray and becoming an entrepreneur by Graham Finlayson. At this point in the week, I was starting to become a little overwhelmed by the amount of information I had taken in on a range of highly technical topics. I think that it could have been more beneficial to have a slightly less dense schedule of lectures and mix in some workshops or seminars to fully take in all of the presentations. Despite this, I did find a lot of value in the lectures on this day, particularly the unsupervised learning lecture in the morning. The evening social event was a relaxed computer vision pub quiz, with a mix of themed questions about computer vision, AI, local and general knowledge. This was again a good time to get to know the other attendees and I thoroughly enjoyed it despite missing out on first place by a couple of points (I blame that on the winning team having a local).

Friday morning consisted of the last couple of lectures of the event. The first one, Art and AI by Chrisantha Fernando being particularly insightful and perhaps my favourite of the week. This lecture, by a Deepmind researcher looked at the state-of-the-art generation models such as Dall-E and asked whether an AI could actually create something more than a picture but what we would consider real “Art”. To examine this idea the speaker then dissected what we mean by the term “art” and emotions using computational terms and ideas and discussed the possibility of AI art through this viewpoint. I found the mix of cognitive science, computer science and philosophy to be very engaging as this cross section of AI is where my own passion for the subject lies.

After the event finished at midday, I met some of the speakers, organisers and attendees for lunch to chat and reflect on the week. Overall, I found the summer school very enjoyable, if not a little bit lecture heavy, and would definitely attend again. I came back from the trip eager to try out some of the more intriguing models and architectures discussed and I will also be going back over some of the key slides when they are released.