# Scalable Scientific Machine Learning Lab

> Scalable Scientific Machine Learning Lab


## Posts

- [3 papers accepted at Differentiable Systems and Scientific Machine Learning workshop @EurIPS 2025](https://scalable-sciml-lab.org/news/3-papers-accepted-at-differentiable-systems-and-scientific-machine-learning-workshop-eurips-2025/): We are excited to announce we have 3 workshop papers accepted at the Differentiable Systems and Scientific Machine Learning workshop at EurIPS on December 6, 2025 in Copenhagen! Here are the papers: Hybrid Learning of Transport Equations with Differentiable Neural Solvers from Experimental Data, Arthur Jessop, Mohammed Alsubeihi, Ashwin Kumar Rajagopalan, Ben Moseley Learning Soil Water Retention Components through an End-to-End Differentiable Hybrid Model, Sarem Norouzi, Per Moldrup, Ben Moseley, David Robinson, Dani Or, Budiman Minasny, Morteza Sadeghi, Tobias L. Hohenbrink, Emmanuel Arthur, Mogens H. Greve, Lis W. de Jonge Discovering and modelling dynamics on latent manifolds with neural geodesic […]
- [Dr. Ben Moseley gives workshop on scalable physics-informed neural networks at CWI Amsterdam](https://scalable-sciml-lab.org/news/dr-ben-moseley-gives-workshop-on-scalable-physics-informed-neural-networks-at-cwi-amsterdam/): Dr. Ben Moseley taught students how to design scalable physics-informed neural networks at Centrum Wiskunde & Informatica in Amsterdam during their Autumn School on Scientific Machine Learning and Numerical Methods. The full workshop recording is now on YouTube, and the slides and practical exercises are on GitHub. Here’s what he covered:Session 1: Introduction to scientific machine learning and PINNsSession 2: Accelerating PINNs with JAXSession 3: Accelerating PINNs with domain decomposition and numerical linear algebra.
- [We’re hiring PhD students!](https://scalable-sciml-lab.org/news/were-hiring-phd-students/): Several PhD opportunities are available in our Scalable Scientific Machine Learning Lab at Imperial College London. These projects are eligible for Imperial PhD scholarships (open to both home and overseas students, with application deadlines between Nov–Jan). Current topics include: 🔬 Multi-scale simulation with physics-informed neural networks🧠 Brain ultrasound imaging with diffusion-guided full-waveform inversion🌍 Learning fast and generalizable climate models with neural differential equations🌀 Learning dynamics on manifolds with neural geodesic flows I also welcome new project proposals related to scientific machine learning. Full details and how to apply here.
- [Ardan Suphi visits University of Bern](https://scalable-sciml-lab.org/news/ardan-suphi-visits-university-of-bern/): One of our PhD students, Ardan Suphi, will be visiting the University of Bern to collaborate on improving multispectral imaging of Mars using SciML. He will be working alongside the Planetary Imaging Group to tailor his research of predictive models to support practical applications such as research of dynamic surface processes on Mars. This collaboration aims to mitigate data loss from HiRISE images, using SciML, and to ensure resulting models are useful, accurate and relevant for other researchers. He will be working with Dr Valentin Bickel (Center for Space and Habitability) and will be in Bern from September 2025, returning […]
- [Multi-scale simulation with physics-informed neural networks](https://scalable-sciml-lab.org/projects/multi-scale-simulation-with-physics-informed-neural-networks/): Overview Physics-informed neural networks (PINNs) have emerged as a promising tool for solving differential equations. They have been applied to many scientific problems and a large number of approaches extending their capabilities have been proposed. PINNs work by using a neural network to directly approximate the solution and training it to satisfy the differential equation. However, a well-known limitation of PINNs is that they often struggle when solving equations with highly multi-scale solutions. This is primarily due to the spectral bias inherent in neural networks (their tendency to learn high frequencies much slower than low frequencies) and the complexity of […]
- [SciML-enhanced planetary exploration: advancing lunar and martian imaging](https://scalable-sciml-lab.org/projects/sciml-enhanced-planetary-exploration-advancing-lunar-and-martian-imaging/): Overview Scientists and engineers leading missions like NASA’s Artemis program and future Mars expeditions, along with planetary researchers studying our solar system’s evolution, rely heavily on high-resolution imaging to plan missions and study planetary bodies and moons. However, the orbital cameras capturing these images operate in extreme environments, and challenges like extremely low-light illumination over permanently shadowed regions on the Moon and the ageing and degradation of sensors over Mars hinder the acquisition of clear images. Our goal is to accelerate discoveries in planetary science by designing state-of-the-art scientific machine learning-based image processing tools that can overcome these challenges. We […]
- [Machine learning with geodesic flows](https://scalable-sciml-lab.org/projects/machine-learning-with-geodesic-flows/): Overview Many physical, biological and engineering systems evolve over time according to geometric laws, for example planets follow elliptical orbits shaped by gravity, and fluids swirl along streamlines governed by curvature and vorticity. A central challenge in modern machine learning is to model these dynamics faithfully and efficiently, especially in cases where the underlying structure is nonlinear, high-dimensional, or only indirectly observed. Most existing deep learning approaches to dynamics approximate trajectories in flat, Euclidean spaces and often overlook the underlying geometry of the data. This leads to models that may be expressive but lack interpretability, physical consistency, or generalisation capacity. […]
- [Efficient and differentiable population balance modelling with JAX](https://scalable-sciml-lab.org/projects/efficient-and-differentiable-population-balance-modelling-with-jax/): Overview Population balance equations (PBEs) are used to model the evolution of populations of particles over time, such as in crystallisation processes, chemical reactors, and biological cell growth. Solving these equations are crucial in sectors like pharmaceuticals, where they allow us to optimise manufacturing processes involving crystallisation and shorten drug development timelines. However, traditional PBE solvers face significant challenges: they are computationally intensive, often requiring substantial time to achieve accurate solutions, and they often rely heavily on empirical laws which may lead to inaccurate physics in the model. In this collaborative project with the Purification And Separation Technology Laboratory at […]
- [Extending quantum theories with AI](https://scalable-sciml-lab.org/projects/extending-quantum-theories-with-ai/): Overview Quantum theory is incredibly powerful for predicting the probabilities of what we’ll see in experiments, but it cannot tell us the certain outcome of a single experiment. That is to say, current quantum theory states that the universe is inherently random. Whether or not this is the case, and whether it is possible for quantum theory to predict certain outcomes, has been a topic of intense debate, with figures such as Einstein suggesting the theory might be incomplete. Extensions of quantum theory have been proposed which try to predict the outcomes of experiments deterministically. Such proposals often involve introducing […]
- [Weather and climate modelling with neural differential equations](https://scalable-sciml-lab.org/projects/weather-and-climate-modelling-with-neural-differential-equations/): Overview This is a new direction for the lab – more to come! Team & collaborators
- [So, what is scientific machine learning?](https://scalable-sciml-lab.org/blog/so-what-is-scientific-machine-learning/): Scientific machine learning (SciML) is an interdisciplinary field that merges the power of machine learning with traditional scientific methods. The field develops techniques for scientific research which combine physics-based knowledge, like mathematical equations and models, with data-driven learning techniques, like deep neural networks. These new techniques help us solve complex scientific problems that are computationally intensive, hard to model accurately, or that involve incomplete data. SciML is empowering researchers to tackle the world’s most pressing problems, from dramatically accelerating weather forecasting, to being able to accurately predict the 3D structure of proteins, and helping us design sustainable nuclear fusion reactors. […]
- [Dr. Ben Moseley joins editorial board of new ACM journal on AI for science](https://scalable-sciml-lab.org/news/dr-ben-moseley-joins-editorial-board-of-new-acm-journal-on-ai-for-science/): We’re excited to share that Dr. Ben Moseley has joined the editorial board of the new ACM Transactions on AI for Science (TAIS) as an Associate Editor. With the rapid growth of AI for science, TAIS aims to provide a venue for cross-disciplinary, rigorous, and trustworthy research. One of the key challenges in this field is communicating clearly across scientific domains — ensuring that work is not only technically sound, but reproducible, well-contextualised, and accessible to a broad scientific audience. Ben is looking forward to supporting this effort and helping build a research culture that values clarity, reproducibility, and collaboration.
- [New SciML lab at Imperial College London!](https://scalable-sciml-lab.org/news/new-sciml-lab-at-imperial-college-london/): We are very excited to announce the formation of our new research group, the Scalable Scientific Machine Learning Lab. The group is led by Dr. Ben Moseley and is part of the Department of Earth Science and Engineering at Imperial College London. Our mission is to accelerate scientific research by designing robust, scalable scientific machine learning algorithms and applying them to impactful problems in Earth science, space science, and beyond. We are looking for people who want to change how scientific research is done, through AI-augmented discovery. If you are interested, check out our current positions. We work closely with […]
- [Dr. Ben Moseley joins Imperial](https://scalable-sciml-lab.org/news/dr-ben-moseley-joins-imperial/): Dr. Ben Moseley will be joining Imperial College London as a Lecturer in AI at the Department of Earth Science and Engineering. He will hold an Eric and Wendy Schmidt AI in Science Fellowship at the Imperial I-X Centre and lead the Scalable Scientific Machine Learning Lab.

## Pages

- [Contact](https://scalable-sciml-lab.org/contact/): 🧑 Lab lead: Dr. Ben Moseley ✉️ Email: info@scalable-sciml-lab.org 🗺️ Location: Royal School of Mines, Imperial College London, South Kensington Campus, London SW7 2AZ
- [About](https://scalable-sciml-lab.org/about/): The Scalable Scientific Machine Learning Lab is led by Dr. Ben Moseley and is part of the Department of Earth Science and Engineering at Imperial College London. We accelerate scientific research by designing scientific machine learning (SciML) algorithms and applying them to impactful problems across science. We develop SciML techniques such as physics-informed neural networks, hybrid ML-numerical algorithms, and physics-based computer vision, and use them to accelerate simulations, better extract knowledge from data, discover new physical models, and improve experiment design. We focus on designing SciML algorithms which are 1) robust, designing physically-grounded workflows that generalise well, and 2) scalable, […]
- [News](https://scalable-sciml-lab.org/news/)
- [Work with us](https://scalable-sciml-lab.org/work-with-us/): Jump to:  PhD positions  Postdoc positions  Master’s projects  Visiting PhD students  Collaboration / consultancy PhD positions Imperial PhD scholarships In September 2026 I (Dr. Ben Moseley) will advertise multiple projects for fully funded PhD scholarships starting in September 2027. See here for more details on Imperial PhD scholarships. See here for projects we have previously offered. Other PhD funding sources In general, I am open to supervising PhD projects on scientific machine learning topics if an appropriate funding source can be found, for example through industry funding. If you’re interested in this route, please send me a research proposal outlining your proposed PhD topic (see Apply). Apply […]
- [Publications](https://scalable-sciml-lab.org/publications/): Key: plain language summary Peer-reviewed journal / main conference papers Local Feature Filtering for Scalable and Well-Conditioned Domain-Decomposed Random Feature Methods.Significantly accelerated the training of physics-informed neural networks by using random features and domain decomposition to turn their optimisation problem into a structured least squares problem and proposing a novel preconditioner to accelerate convergence.Willem van Beek, J., Dolean, V., Moseley, B. (2025).Computer Methods in Applied Mechanics and Engineering.Paper    Code    Workshop Challenges and advancements in modeling shock fronts with physics-informed neural networks: A review and benchmarking study.Explores how to improve physics-informed neural networks for solving equations that involve sudden physical changes – like […]
- [Projects](https://scalable-sciml-lab.org/projects/): Tags
- [People](https://scalable-sciml-lab.org/people/): Current Dr. Ben Moseley (Assistant Professor) Group Head Ardan Suphi PhD Student Davide Staub PhD Student Master’s students Amit Chandra Collaborators Dr. Ghislain Fourny Lecturer (ETH Zurich) Dr. Valentin Bickel CSH Fellow (University of Bern) Sarem Norouzi* PhD Student (Aarhus University)*Main supervisor: Prof. Lis Wollesen de Jonge Alumni 2025 Master’s students Emil Sieciechowicz Lewis O’Donnell Philéas Grenade Timothee Boudot Akshay Somvanshi Jiayu Li Mathilde Horanieh 2024 Julian Bürge Master’s Student (ETH Zurich) [Thesis] Florian Pauschitz* Master’s Student (ETH Zurich) *Co-supervision with Ghislain Fourny [Thesis] Sihan Cao Master’s Student


[comment]: # (Generated by Hostinger Tools Plugin)