Intelligent Earth (UKRI CDT in AI for the Environment)

This course will introduce the essential tools and best practices needed for a sustainable approach towards software engineering in a Python programming environment. This will include an overview of programming paradigms (eg object orientated, functional and array-based programming), best practices in construction and design, the use of code repositories, version control, integrated development environments, automated test-driven development, debugging, continuous integration, and deployment and packaging of code-based tools as well as agentic programming. 

This course will introduce you to common Python data analysis tools and data formats used in the environmental sciences. We will cover the basics of loading, selecting, plotting, and performing operations on different data structures, with interactive notebooks and examples for you to work through and consolidate your learning.

Statistical inference encompasses the estimation of parameters, determining the uncertainty associated with those parameters, and making categorical statements about them. While disciplines like applied mathematics and engineering also involve estimating quantities and making decisions based on data, statistics places emphasis on measuring uncertainty in those estimates and decisions. This course aims to provide solid foundations to implementing linear regression, generalised linear Models, and hierarchical models.

Machine learning is a collection of techniques that allow creation or tuning of models using example data or past experience. This course provides an introduction to machine learning with an emphasis on the recent research and application landscape. The course introduces fundamental techniques including differentiable programming and generative modeling. The course also covers key concepts in the actual application of machine learning to practical problems, such as data pre-processing and model evaluation.

In this course we provide the background knowledge needed to understand the physical climate system. We take you from fundamental ideas in radiation thermodynamics and fluid dynamics through to complex emergent phenomena such as the atmospheric and oceanic circulations. We discuss how we observe the current climate system and introduce the models (both conceptual and computational) that we use to make predictions of the future state of this system. Throughout, a focus is on the potential for Machine Learning to address problems in climate science, including examples of where its transformative power is already being felt.

This course provides an introduction to natural hazards experienced on Earth. The first two days of the course will introduce you to natural hazards arising from processes at work in the solid Earth and cryosphere, including seismic hazard, volcanic eruptions, and sea-level rise. The practical sessions in these first two days will introduce the wide variety of data types used in the observation of solid-Earth and cryosphere processes, as well as emerging AI/ML approaches for analysing these data types. The final two days of the course will introduce you to a range of water and weather-related extremes, including floods, droughts, extreme heat/heatwaves, intense rainfall and storms (including convective-scale modelling), and storm surges/coastal risks.

This course will describe the generation, maintenance, and loss of biodiversity, along with key challenges in this area. Lectures will give an overview of current understanding and will point you to further, more specialist, information. The practical sessions will introduce you to some large data sets characterising biodiversity and you will interrogate these data with AI and ML methods to address outstanding questions.

This course introduces four Environmental Solutions themes: nature recovery, renewable energy, carbon removal and net zero, and adaptation and infrastructure. You will explore how AI and data-driven methods can support practical environmental action, from monitoring ecosystems and clean energy systems to assessing emissions pathways and climate-resilient infrastructure. The week links core environmental challenges with examples of how quantitative methods can inform decision-making and implementation.

This course on Responsible AI for the Environment will introduce you to the key concepts of responsible AI, such as fairness, inclusivity, reliability, privacy, trustworthiness, transparency and accountability, in the context of environmental research. It will also serve as a guide for you throughout your DPhil.

You will be introduced to the key concepts of Responsible AI and AI governance, followed by group work to assess Responsible AI considerations across each of the IE themes. We will draft a whitepaper on Responsible AI for the Environment as well as a responsible AI checklist that will also guide your own research

Signal processing and time series analysis are crucial for extracting valuable information from data that changes over time. These techniques allow us to understand patterns, make predictions, and gain insights from signals. Time series modelling underpins sensor data analysis for monitoring environmental conditions, detecting anomalies and tracking activity; it enables ecological monitoring using bioacoustics and is the foundation for much decision and detection in the data we collect.  

This course will introduce the foundational concepts of dealing with time varying data, how to filter it, represent it and extract information from it. We consider conventional time series data and signals that occur in networks of interconnected devices, so requiring methods from graph signal processing. The course ends by merging time series methods with machine learning, considering how data can be forecast and events identified.

This deep learning course provides a comprehensive introduction to the fundamentals of machine learning workflows and key concepts, progressing to advanced topics and real-world applications. 

Week one focuses on foundational topics like training, inference, transfer learning, and briefly computer vision, covering architectures such as CNNs and Vision Transformers. 

Week two explores advanced themes, including fine-tuning large language models, reinforcement learning, geometric deep learning, and explainable AI.

Through a combination of theoretical instruction and practical lab sessions, you will gain the skills needed to implement deep learning models in diverse domains like vision, language, and climate science.

This course provides a comprehensive overview of computer vision, covering key aspects of image formation, 2D vision, and 3D vision. 

It begins with the fundamentals of geometric projections, transformations, and photometric principles in image acquisition. You will explore 2D vision techniques such as feature extraction, matching, and semantic understanding, including object recognition, detection, and segmentation. 

The course also delves into 3D vision, focusing on camera pose estimation, depth triangulation, and 3D reconstruction, equipping you with the skills to apply these concepts in real-world applications like autonomous driving, virtual reality, and 3D modelling.

Examples of professional skills training include:

• Public Engagement: public engagement with research is an important part of being a researcher and a way to connect society and research: by sharing what research is being done, and involving the public in the design and conduct of research to ensure it’s relevant to society’s needs and interests. Public engagement with research comes in many forms and can serve many purposes. A typical form of engagement is to inspire and inform local public audiences through appealing, interactive, hands-on experiences.
• Publishing in Environmental AI: To help you to navigate the complex landscape of interdisciplinary publishing between environmental science (primarily journal based) and AI (primarily conference based). 

A secondment with non-academic partners may take place in year three (years five to six if studying part-time), if it was not undertaken previously.

In year three (years five to six if studying part-time), alongside your research you will have the opportunity to take the following, based on your individual needs:

• Optional advanced courses: further courses to focus on specific areas of AI applied to grand challenges and associated datasets from the four environmental themes
• Professional skills training modules
• Getting started in policy engagement
• Planning to write a DPhil
• Developing your employability

The focus of Intelligent Earth is AI for the environment. Themes of this course include:

• Climate
• Biodiversity
• Natural hazards
• Environmental solutions
• Core AI/ML research on complex environmental data.

In year one (first two years if studying part-time) you will undertake a four-month research project or two two-month projects, followed by the annual hackathon and conference. The teaching model for all courses will be tailored towards training students to become independent researchers.

Information about academic and research staff can be found on the centre's website. 

Official course title

Intelligent Earth (UK Research and Innovation Centre for Doctoral Training (CDT) in Artificial Intelligence for the Environment) 

Degree award

Course code

Full time: R30_1

Part time: R30_9P1

The full-time course is usually studied over three to four years. The part-time course is usually studied over six to eight years.

If you are studying part-time you will be required to commit to allocating an average of 20 hours per week (50% full time equivalent) to doctoral study. As a part-time student you will follow an individually-tailored training programme that is adjusted to suit your work pattern, with the opportunity to undertake additional training as needed during the later years of your studies. Where possible you are encouraged to allocate additional time to your studies at the beginning of your first year of study to fully benefit from induction processes, cohort interactions and foundational training. If you are part-time you will be assessed at milestones that are adjusted to take your working hours into account (eg 12 terms for transfer of status for a student working at 50% FTE).

One of the following departments will serve as your primary department from year two onwards, when you will begin your DPhil research:

• Department of Biology
• Department of Computer Science
• Department of Earth Sciences
• Department of Engineering Science
• School of Geography and the Environment
• Department of Physics
• Department of Statistics

The Intelligent Earth CDT is resourced by UK Research and Innovation (UKRI) with contributions from the University and partners. It is integrated in Oxford's Doctoral Training Centre and training may take place in a number of locations across the University. Students will often work together on problem sets, or in groups, with the support of demonstrators.

You will have access to computational facilities as well as a laptop. When you move out to your department you will also have access to the facilities provided by that department. You will remain a member of the CDT and be able to return to the CDT facilities, based within the Doctoral Training Centre, on Keble Road, to use the facilities there.

You will have access to seminars in all seven departments as well as across the wider university. In addition to the training modules offered by the CDT, you will be able to sign up for a wide range of training courses and modules offered by departments across the university via the University's Researcher Training Tool.

You will also have access to Oxford's wide library network, including the recently refurbished Radcliffe Science Library.

The allocation of graduate supervision for this course is the responsibility of the Intelligent Earth CDT and it is not always possible to accommodate the preferences of incoming graduate students to work with a particular member of staff. Under exceptional circumstances a supervisor may be found outside the Intelligent Earth CDT.

During your first year (first two years if studying part-time), you will have a supervisor from the academic leadership of the CDT. Meetings with your supervisor will serve to monitor academic progress as well as to discuss any academic issues or questions arising. When you transition into one of the participating departments to commence your research project in year two (three to four if studying part-time), you will be co-advised by two supervisors, one from an environmental science department and one from an AI department. First-year supervisors will act as mentors throughout the programme, providing academic and pastoral guidance.

Students will have the opportunity to meet their supervisors on a regular basis to discuss their progress. For full- time students, these meeting are usually at least once every two weeks, averaged across the year, and typically once a month for part-time students.

You will be assessed continually throughout the first-year training courses modules. In the second half you will undertake your research project(s) and will be required to deliver a written report that will be assessed.

At the end of the second year, you will be required to write a report and give a presentation on your research, and to present a detailed and coherent plan for the research-intensive phase in the third and fourth years of your doctoral studies. Progress towards completion is again formally assessed some way into the final year of study.

If studying part-time, you will be assessed over the first two years. Timings will differ for the presentation for part-time students.

You will carry out your DPhil project in the department of your primary supervisor and will gain your DPhil from the department in which you carry out your research project.

You will follow the same milestones and assessments as a standard DPhil. All students will be initially admitted to the status of Probationer Research Student (PRS).  The first milestone is your transfer of status from PRS to DPhil student. The second milestone is the confirmation of status. These milestones are typically undertaken in the second and third year of your degree respectively. The exact assessment format used will vary according to discipline and department, but each typically involves an assessment of your progress by two researchers with relevant expertise. If studying part-time the milestones will differ to a full-time student. 

At the end of your studies you will submit and be formally assessed on a written thesis describing your research. Specific thesis requirements are set by the department in which you are registered to undertake your substantive DPhil project. To be successfully awarded a DPhil you will need to defend your thesis orally (viva voce) in front of two appointed examiners.

The CDT will train a new generation of quantitative environmental data scientists equipped to make substantial contributions in environmental and data sciences as well as being prepared for a wide range of career paths in academia, research and industry, supported by the CDT's extensive partnership network.

The CDT anticipates Intelligent Earth Graduates to drive innovation and found their own start-ups, supported by the programme’s dedicated training in enterprise, impact, and responsible AI.

This course is led by the MPLS Doctoral Training Centre

• Course page on the Intelligent Earth website
• Funding information from the CDT
• Academic and research staff at the CDT
• MPLS Doctoral Training Centre website
• Mathematical, Physical and Life Sciences
• Contact details

*Combined figure for full-time and part-time study.
^†Three-year average (applications for entry in 2024-25 to 2026-27)
^ Applications for entry in 2026-27 (first year of admissions) 

• a first-class or strong upper second-class undergraduate degree with honours

The above qualification should be achieved in one of the following subject areas or disciplines:

• biology
• chemistry
• computer science
• engineering
• earth sciences
• environment sciences
• geography
• geology
• geophysics
• mathematics
• meteorology
• oceanography
• physics
• statistics; or
• other related disciplines such as data science, mathematical, physical and environmental science

For applicants with a bachelor's degree from the USA, the minimum overall GPA that is normally required to meet the undergraduate-level requirement is 3.5 out of 4.0.

GRE General Test scores

No Graduate Record Examination (GRE) or GMAT scores are sought.

Other qualifications, evidence of excellence and relevant experience

• The course aims to develop and apply AI and candidates should have strong quantitative skills. Therefore, as a minimum applicants should have maths skills equivalent to UK A-level or to have done maths at an equivalent level as part of their course. If not obvious from your submitted documents, please provide brief details in your statement.
• Candidates will need to demonstrate a broad interest in some or all of the five Intelligent Earth themes: Climate, Biodiversity, Natural Hazards, Environmental Solutions, and Core AI Research in their personal statement and interview.
• Applicants whose previous degree was not explicitly quantitatively based should be able to demonstrate the ability to learn the necessary skills required for successful completion of the course.
• Although it is not required to have a master’s degree, in practice most applicants with a ‘physical science’ background will have completed a four-year integrated master's course.
• Professional experience, especially research experience, is valuable and will be taken into consideration as a substitute for an academic qualification.
• Publications are not expected, but should be included if present.

Interviews are normally held as part of the admissions process and are expected to be held in the second half of February.

Further information about funding opportunities for this course can be found on the Intelligent Earth CDT website. 

Full-time study

Part-time study

Full-time study

Part-time study

Full time

The following colleges accept students for full-time study on this course:

• Balliol College
• Brasenose College
• Campion Hall
• Exeter College
• Green Templeton College
• Hertford College
• Jesus College
• Keble College
• Lady Margaret Hall
• Linacre College
• Lincoln College
• Mansfield College
• New College
• Pembroke College
• Reuben College
• St Anne's College
• St Catherine's College
• St Cross College
• St John's College
• University College
• Wadham College
• Wolfson College
• Worcester College
• Wycliffe Hall

Part time

The following colleges accept students for part-time study on this course:

• Exeter College
• Keble College
• Wadham College
• Wolfson College
• Worcester College

Informal enquiries should be made to the Intelligent Earth CDT administrator in the first instance.

You are encouraged to communicate with academics working in your area of interest to discuss potential research topics and the possibility of being offered supervision but this does not affect the admissions process. Profiles of academics with whom you might wish to study can be found on the CDT website.

If you are not applying for a specific project, you do not have to specify one at this stage. Instead, you should indicate your preferred CDT Research theme under 'Proposed field and title of research project'. If there are pre-defined projects or topic areas of interest to you, you can indicate this in your statement of purpose. 

It is not necessary for you to identify a potential supervisor in your application.

Students will be matched with supervisors and partners throughout the first year of training. Nonetheless, you are encouraged to make contact with potential supervisors throughout the application process. This may lead to the development of a project idea, but there is no commitment on either side, nor does it affect the admissions process.

Your statement should be written in English and should focus on your motivation, research interests and career ambitions in the area of the CDT, rather than on other personal achievements, interests and aspirations. It should refer directly and specifically to one or more of the themes of the CDT.

It will be normal for your ideas and goals to change in some ways as you participate in the programme and you are not committed to work in the specific subject area. You should nevertheless make the best effort to demonstrate your current interests and aspirations. 

Your statement will be assessed for:

• your reasons for applying
• your ability to present a coherent case
• your commitment to the subject, beyond the requirements of the degree course
• your preliminary knowledge of the subject area and research techniques
• your reasoning ability
• your ability to absorb new ideas

Email: [email protected]