PhD Position Physics-Aware Foundation Models for Climate Research

The emergence of general-purpose large AI models, known as "foundation models" (FMs), such as GPT-n and Llama for language, and DALL-E for images, marks a significant advancement in AI. However, their application to critical areas like climate, environment, and natural energy often lacks interpretability and reliability, largely due to their data-driven nature. Integrating principles of physical consistency into FMs is crucial to ensure that the models not only perform well on data but also adhere to the fundamental principles of the systems they aim to model.
This PhD project will address the following fundamental challenges on how to achieve this for practical large-scale systems.:

• Developing a large-scale physically consistent FM: You will integrate atmosphere, surface, and subsurface earth observations to develop FMs for modelling the erath and climate system that incorporate multiple physical principles, especially considering system dynamics. You will explore incorportating prior physics knowledge throughg the use of datasets, loss functions, or model architectures.
• Balancing interpretability and generalization: While adding physical constraints can improve interpretability and reliability, it can also potentially restrict the model's ability to generalize to diverse datasets or adapt to novel scenarios. You will investigate balancing between incorporating constraints and maintaining flexibility from a more fundamental aspect.
• Enhancing Sustainability: You will investigate how to reduce the energy consumption associated with training and maintaining FMs, addressing urgent climate challenges.

Requirements:

To be considered for the position, you will have:

• MSc degree in Computer Science, Artificial Intelligence, Data Science, Geophysics, Applied Physics, Computational Physics, or a closely related field.
• Good theoretical understanding of the fundamentals of Machine and Deep Learning.
• Basic knowledge and a keen interest in Applied (Geo)Physics.
• Strong programming skills (preferably Python) and experience with Machine and Deep Learning tools and techniques.
• Ability to work independently (taking initiative, being organized and creative) as well as to collaborate effectively across multiple disciplines (especially with geophysicists).
• Strong interpersonal communication and collaboration abilities
• Good communication skills and proficiency in English.

Doing a PhD at TU Delft requires English proficiency at a certain level to ensure that the candidate is able to communicate and interact well, participate in English-taught Doctoral Education courses, and write scientific articles and a final thesis. For more details please check the Graduate Schools Admission Requirements.

Salary Benefits:

Doctoral candidates will be offered a 4-year period of employment in principle, but in the form of 2 employment contracts. An initial 1,5 year contract with an official go/no go progress assessment within 15 months. Followed by an additional contract for the remaining 2,5 years assuming everything goes well and performance requirements are met.
Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities, increasing from € 2872 per month in the first year to € 3670 in the fourth year. As a PhD candidate you will be enrolled in the TU Delft Graduate School. The TU Delft Graduate School provides an inspiring research environment with an excellent team of supervisors, academic staff and a mentor. The Doctoral Education Programme is aimed at developing your transferable, discipline-related and research skills.
The TU Delft offers a customisable compensation package, discounts on health insurance, and a monthly work costs contribution. Flexible work schedules can be arranged.
For international applicants, TU Delft has the Coming to Delft Service. This service provides information for new international employees to help you prepare the relocation and to settle in the Netherlands. The Coming to Delft Service offers a Dual Career Programme for partners and they organise events to expand your (social) network.

36 - 40 hours per week

Mekelweg 2