Postgrad LIVE! Study Fairs

Birmingham | Edinburgh | Liverpool | Sheffield | Southampton | Bristol

University of Bristol Featured PhD Programmes
University of Edinburgh Featured PhD Programmes
University of Portsmouth Featured PhD Programmes
University of Oxford Featured PhD Programmes
University of Manchester Featured PhD Programmes

DeepLandscape: Integrating Artificial Intelligence into practices of archaeological landscape interpretation

This project is no longer listed in the FindAPhD
database and may not be available.

Click here to search the FindAPhD database
for PhD studentship opportunities
  • Full or part time
    Dr R Opitz
    Dr JP Siebert
  • Application Deadline
    No more applications being accepted
  • Competition Funded PhD Project (European/UK Students Only)
    Competition Funded PhD Project (European/UK Students Only)

Project Description

This PhD project aims to develop a method for integrating AI-led survey with contemporary topographic interpretation practices, and to reflect on the impact of the introduction of this new approach on professional practice. In the context of landscape archaeology and topographic interpretation, in order to take advantage of AI-led approaches, e.g. the automatic identification of features and changes, we must develop a framework for integrating AI into practices of archaeological landscape interpretation, a practice that is currently entirely based on individual visual observation. This requires a critical examination of our interpretive practices, interrogating the influence of our individual experiences, and asking how new interpretive practices might be developed.

The PhD project will take advantage of a pilot project on Arran, where AI-based automated identification of several classes of the islands archaeological features has been conducted within the framework of an archaeological survey of the island by Historic Environment Scotland (HES). The pilot study suggests that creating sound training sets for the identification of archaeological features in lidar topographic models is critical to the process of integrating AI-led approaches and contemporary survey practices. This task is particularly challenging because the identifications of training features are inherently highly interpretive and prejudiced by (expert) observers’ prior knowledge. To mitigate this, we must ask: How can we understand the relationship between what observers identify and what the AI identifies?

The project will have five main stages to study the relationship between human and AI led visual interpretations of topography: application of next generation DCNN methods: an exploration of deep net architectures and combinations of learned and defined features, and integration of multi-modal data-sets; method development for archaeologists’ fieldwork using AI-identifications; integration of archaeologists’ feedback; method development for archaeologists’ creating training data, and reflection on the impacts of AI methods on practice. Issues considered will include agreement between identifications by the deep net and archaeological professionals pre- and post- the addition of new training data, and responses to pre- and post- workshop surveys of archaeological professionals.

This project is embedded at HES and consequently will have direct impact on practice in Scotland. It uses lidar as its test data type, but the principles of the approach are readily expanded to HES’ archives of historic aerial photographic and more recent multi-temporal spectral dataset acquisitions and a variety of datasets increasingly produced through precision agriculture. The use of all these resources is limited by the shortage of expert human interpreters. This problem is critical, as national and international agencies responsible for the management of heritage and individual archaeological researchers working on a variety of questions from how ceramics represent trade to the role of charcoal production in the rural economy are looking to adopt AI led approaches to scale up their studies and address landscapes and assemblages as a whole, rather than through limited case studies.

The project will be formally co-supervised by Dr Rachel Opitz and Dr Jan Paul Siebert (University of Glasgow) and will include regular meetings and further supervision by Mr Dave Cowley (HES). The student will therefore be a member of both the Computer Science and Archaeology postgraduate communities at University of Glasgow and engage directly with the archaeological community at HES.

Applicants should submit a Curriculum Vitae, including contact details of one academic referee, and a 2-page covering letter outlining why they are interested in this collaborative doctoral award and what they would bring to this project.

This should be sent in an email to [Email Address Removed] and [Email Address Removed] by 14 December 2018.

Interviews will be held on 16 January 2019. This will enable the identification of a candidate who will liaise with the supervisory team and complete a full CDA PhD studentship application form by 13th February 2019, for consideration and final evaluation by SGSAH. Those successfully nominated will not be automatically funded.

Funding Notes

To be eligible for a full award a student must have a relevant connection with the UK:
• The candidate has been ordinarily resident in the UK, meaning they have no restrictions on how long they can stay
• Been ‘ordinarily resident’ in the UK for 3 years prior to start date
• Not been residing in UK wholly or mainly for the purpose of full-time education. (Not applicable to UK/EU nationals).
Students from EU countries other than UK are generally eligible for a fees-only award. These students must be ordinarily resident in a member state of the EU.

FindAPhD. Copyright 2005-2018
All rights reserved.