The following questions are aimed to capture a snapshot of an individual student within one the CDT cohorts. During proposal development, hypothetical students (with real Southampton supervisors, and hypothetical industrial ones) are used to help shape the training programme and give light to how the CDT would be experienced from the perspective of the student. If successfully funded, the same questions will be reused to help us tell the stories of our actual students.

This notional case study is based around a previous Southampton student, Dr Iris Kramer, who is now actually involved in the CDT development through her company ArchAI , which is one of our partners. This case study has allowed us to reflect on how Iris’s own experience would have been different if she had had the opportunity to be part of the CDT.

Describe the student’s previous academic and/or professional background

XXX undertook a Batchelor’s degree in Archaeology in Europe, and an MSc in Archaeological Computing (GIS and Survey) in Southampton. She also undertook a short course in software development to bring up her programming skills and knowledge.

What is the student’s motivation for joining the Geospatial AI CDT at Southampton?

After her MSc, XXX became really interested in the possibility of using AI techniques to automate the detection of archaeological sites from geospatial data including aerial imagery and airborne LIDAR. The CDT offered her exactly the opportunity she needed to undertake the training and develop the skills (in both applied machine learning technologies and translating research into a new business) required to exploit this idea. In addition, the CDT’s partners offered access to the data she needed to complete her research idea and access to internships where she could work closely with industry professionals interested in related ideas. The CDT’s enterprise training, and the opportunities presented by the close involvement of FutureWorlds was also appealing as XXX could envisage ways in which her research idea might be commercialised.

Who are the student’s supervisors? What is their research topic/area?

XXX was supervised by Professor Jonathon Hare and Professor Adam Prügel-Bennett in ECS, and Dr Isabel Sargent from Ordnance Survey (who contributed to the studentship financially and provided data). The supervisory team has a broad interest in representation learning and self-supervised approaches. Isabel and Jonathon had previously collaborated on projects looking at using deep learning to extract geospatial information from aerial photographs.

How has the Geospatial AI CDT provided a technical training programme that is tailored to the student’s needs and their research topic/area?

The CDT provided a tailored programme to help XXX achieve her goals. XXX had considerable domain knowledge from her previous degrees, but the CDT provided ground-up training in numerical computing, geospatial data manipulation using the Python programming language and deep learning technologies. XXX also took advantage of the enterprise training offerings provided by the CDT to allow her to continue taking forward her research idea towards a commercial product after her time in the CDT was over.

How has the Geospatial AI CDT provided training required to allow the student to become a future leader in the field? Are there any particular transferrable or generic skills that this student has mastered?

The enterprise training provided by the CDT gave XXX the grounding and confidence needed to launch a start-up as a result of her research during her time in the CDT. Training in generic skills, such as presentation and negotiation, has allowed her to win funding from government and VCs to help her grow her start-up. Within her growing team in the start-up, she has less time to stay on top of the latest technical developments in Deep Learning for Geospatial Intelligence, but the training the CDT provided allows her to work closely with her technical team. Training and exposure in talking to and working with governmental policy makers has helped XXX develop a network through which her startup is acquiring significant income.

In what ways has this student excelled in working in a group within their cohort during the training programme? How has the training programme facilitated this?

XXX’s background was unique within her cohort; she was the only one with an archelogy background. XXX had natural ability to bring together the group when they were working together on hackathon projects during within CDT programme. Crucially however the training programme gave her the ability of understand different terminologies and diverse viewpoints at a technical level and help the group reason about the best approach to take together. Additionally, the training programme has helped XXX become much more adept at describing the same problem to different groups of stakeholders.

Has the student generated any intellectual property during the programme? (Briefly) what is it? Do the findings impact the public, private or third/voluntary sectors, and elsewhere?

The PhD research became the commercial basis for XXX’s start-up company.

Which CDT partner(s) are (co-)sponsoring the studentship award?

Ordnance Survey co-sponsored the studentship and provided in-kind support through staff time for supervision, and through data access.

If the student has undertaken the internship who was it with, and what did they do?

XXX spent some time working with Ordnance Survey at their offices. The work looked at ways in which the findings from the PhD research (focussed on detecting archaeology) could be translated to other domains.

If the student has finished, where are they going next?

XXX has launched a start-up based on their research and also secured a RAEng Enterprise Fellowship.

How have Southampton’s unique facilities and broad range of Geospatial and AI expertise benefitted the student?

Geographical closeness to the sponsor was a big part, but the broad range of expertise across the different faculties and schools of the University ensured that there was always someone to speak to about any problem. In terms of facilities, the computing infrastructure for working with large, high dimentional, geospatial image datasets provided by the CDT through the host School and wider University was critical to facilitating the research.