PhD Multimodal Generative AI for Uncertainty-Aware 3D Geological Modelling

Project Description

You will develop machine-learning methods to extract and combine information from surface geological maps, cross-sections and well logs while preserving their spatial geometry, scale and resolution. You will also investigate how language models can translate geological descriptions and technical reports into structured and traceable information about geological units, faults and stratigraphic relationships.

The project will explore advanced generative approaches, including conditional diffusion and flow-matching models, for constructing 3D geological structures and lithological volumes. Gravity, magnetic and selected seismic observations will provide additional physical constraints. Differentiable geophysical forward models or machine-learning surrogates will be used to compare the predicted responses of generated models with measured data and guide the generation process towards interpretations that are consistent with observations.

You will create synthetic geological datasets for model training and controlled benchmarking before applying the methods to an industry-supported case study. The research will evaluate geological realism, consistency with observations, ensemble diversity, computational efficiency and uncertainty calibration. You will also investigate the value of different data sources by determining which observations, individually or in combination, reduce geological uncertainty most effectively.

The project is supported by Oxus Metals, which will contribute datasets, an industry case study and specialist geological expertise. Working with the industry partner, you will assess whether the generated models are geologically plausible, whether the underlying assumptions are appropriate and how the resulting methods could support practical geological interpretation and decision-making.

You will join the Institute of GeoEnergy Engineering at Heriot-Watt University and receive interdisciplinary training in geological modelling, geophysical data integration, generative AI, scientific programming and uncertainty quantification. You will have opportunities to publish your research, present at national and international conferences, and collaborate with academic and industry researchers.

The project will deliver a reproducible modelling framework and an auditable prototype capable of generating 3D geological model ensembles, voxel-scale lithology probabilities and diagnostic maps identifying where observations are poorly explained. These capabilities could support applications including critical-mineral exploration, geological CO₂ storage, geothermal energy, underground hydrogen storage, groundwater management and subsurface infrastructure. The combination of geoscience, artificial intelligence and industrial experience will prepare you for careers in the energy and minerals sectors, scientific computing, environmental consultancy or academic research.

Eligibility:

This project is open to candidates who are eligible for home (UK) fee status only.

Applicants should hold, or expect to obtain, at least a 2:1 degree in geology, geophysics, Earth science, geoenergy or petroleum engineering, civil or environmental engineering, computer science, data science, applied mathematics, or a closely related discipline. A relevant master’s degree is desirable but not essential.

Candidates should demonstrate strong quantitative and analytical ability, together with experience in programming, preferably Python. Previous experience in machine learning, geological modelling, geophysical data analysis, numerical modelling or uncertainty quantification would be advantageous. Applicants from either geoscience or computational backgrounds are welcome, provided they demonstrate a strong interest in interdisciplinary research at the interface of geology and artificial intelligence.

Funding

This project is funded by an EPSRC scholarship which will cover tuition fees at the Home rate and provide a stipend (£21,805 in 2026-27) for 42 months. Thereafter, candidates will be expected to pay a continuing affiliation fee (currently £130) whilst they complete writing up their thesis.

How to Apply

1. To apply you must complete our online application form.
2. Please select PhD GeoEnergy Engineering as the programme and include the full project title, reference number (EGIS2026DLA-NW) and supervisor name on your application form. Ensure that all fields marked as ‘required’ are complete.
3. Once have entered your personal details, click submit. You will be asked to upload your supporting documents. You must complete the section marked project proposal; provide a supporting statement (1-2 A4 pages) documenting your reasons for applying to this particular project, outlining your suitability and how you would approach the project. You must also upload your CV, a copy of your degree certificate and relevant transcripts and an academic reference in the relevant section of the application form.
4. If your first language is not English, we'll need to see evidence of your English language ability. The minimum English language requirement for entry to this programme is IELTS 6.5 (or equivalent) with no score lower than 6.0.If you do not have IELTS 6.5, we offer a range of English language courses to help you meet the English language requirement for this programme prior to commencing your studies. For more information about your application and our English Language requirements, please see Section 10 of our page on English Language Requirements as part of your application.

Enquiries:

Please contact Dr Nanzhe Wang (Nanzhe.Wang@hw.ac.uk) for further information or an informal discussion.

Please contact egis-pgr-apps@hw.ac.uk for technical support with your application.

Timeline:

Applications will remain open until the position is filled and will be reviewed every two weeks, beginning Friday 13 November 2026. Interviews will be arranged promptly after each review point, so early application is strongly encouraged. The successful candidate must be identified by 31 January 2027. The start date will be agreed with the successful candidate and may be between January and October 2027.