skip to content

Microstructural Kinetics Group

Department of Materials Science & Metallurgy
 

Wed 12 Feb 13:00: Short-term, high-resolution sea ice forecasting with diffusion model ensembles

School of Physical Sciences - Mon, 10/02/2025 - 15:27
Short-term, high-resolution sea ice forecasting with diffusion model ensembles

Sea ice plays a key role in Earth’s climate system and exhibits significant seasonal variability as it advances and retreats across the Arctic and Antarctic every year. The production of sea ice forecasts provides great scientific and practical value to stakeholders across the polar regions, informing shipping, conservation, logistics, and the daily lives of inhabitants of local communities. Machine learning offers a promising means by which to develop such forecasts, capturing the nonlinear dynamics and subtle spatiotemporal patterns at play as effectively—if not more effectively—than conventional physics-based models. In particular, the ability of deep generative models to produce probabilistic forecasts which acknowledge the inherent stochasticity of sea ice processes and represent uncertainty by design make them a sensible choice for the task of sea ice forecasting. Diffusion models, a class of deep generative models, present a strong option given their state-of-the-art performance on computer vision tasks and their strong track record when adapted to spatiotemporal modelling tasks in weather and climate domains. In this talk, I will present preliminary results from a IceNet-like [1] diffusion model trained to autoregressively forecast daily, 6.25 km resolution sea ice concentration in the Bellingshausen Sea along the Antarctic Peninsula. I will also touch on the downstream applications for these forecasts, from conservation to marine route planning, which are under development at the British Antarctic Survey (BAS). I welcome ideas and suggestions for improvement and look forward to discussing opportunities for collaboration within and beyond BAS .

[1] Andersson, Tom R., et al. “Seasonal Arctic sea ice forecasting with probabilistic deep learning.” Nature communications 12.1 (2021): 5124. https://www.nature.com/articles/s41467-021-25257-4

Add to your calendar or Include in your list

Thu 13 Mar 16:00: Algorithmic stability for regression and classification

School of Physical Sciences - Mon, 10/02/2025 - 14:41
Algorithmic stability for regression and classification

In a supervised learning setting, a model fitting algorithm is unstable if small perturbations to the input (the training data) can often lead to large perturbations in the output (say, predictions returned by the fitted model). Algorithmic stability is a desirable property with many important implications such as generalization and robustness, but testing the stability property empirically is known to be impossible in the setting of complex black-box models. In this work, we establish that bagging any black-box regression algorithm automatically ensures that stability holds, with no assumptions on the algorithm or the data. Furthermore, we construct a new framework for defining stability in the context of classification, and show that using bagging to estimate our uncertainty about the output label will again allow stability guarantees for any black-box model. This work is joint with Jake Soloff and Rebecca Willett.

Evaluating a black-box algorithm: stability, risk, and model comparisons

When we run a complex algorithm on real data, it is standard to use a holdout set, or a cross-validation strategy, to evaluate its behavior and performance. When we do so, are we learning information about the algorithm itself, or only about the particular fitted model(s) that this particular data set produced? In this talk, we will establish fundamental hardness results on the problem of empirically evaluating properties of a black-box algorithm, such as its stability and its average risk, in the distribution-free setting. This work is joint with Yuetian Luo and Byol Kim.

A wine reception in the Central Core will follow this lecture

Add to your calendar or Include in your list

Fri 14 Feb 16:00: Flavour Physics at the Intensity Frontier

School of Physical Sciences - Mon, 10/02/2025 - 14:14
Flavour Physics at the Intensity Frontier

Flavour physics and CP violation provide powerful probes for physics beyond the Standard Model that are sensitive to very high energy scales beyond direct detection experiments. The construction of the unitarity triangle provides the ultimate probe for flavoured new physics. We discuss current highlights of this program in the sectors of beauty, charm and strange quark decays.

Add to your calendar or Include in your list

Wed 12 Feb 16:30: Tensor product functoriality via p-adic propagation

School of Physical Sciences - Mon, 10/02/2025 - 13:29
Tensor product functoriality via p-adic propagation

Tensor product functoriality predicts that the tensor product of two automorphic Galois representations should be automorphic. We will motivate this statement, and describe forthcoming work establishing this conjecture in some cases.

Add to your calendar or Include in your list

Fri 21 Feb 16:00: Title to be confirmed

School of Physical Sciences - Mon, 10/02/2025 - 13:04
Title to be confirmed

Abstract not available

Add to your calendar or Include in your list

Fri 21 Mar 16:00: Title to be confirmed

School of Physical Sciences - Mon, 10/02/2025 - 13:04
Title to be confirmed

Abstract not available

Add to your calendar or Include in your list

Wed 19 Feb 14:30: Unraveling Water’s Behavior in Anisotropic Environments

School of Physical Sciences - Mon, 10/02/2025 - 12:02
Unraveling Water’s Behavior in Anisotropic Environments

The structure and reactivity of water in anisotropic environments, such as at interfaces or under the influence of electric fields, can differ significantly from those observed in bulk. Understanding these differences is crucial for gaining insights into various atmospheric and electrochemical processes that impact our society.

In the first part of my talk, I will discuss how ions organize at the water/air interface and demonstrate that the conventional electric double-layer model fails to provide a complete microscopic picture of these interfaces [1,2]. Using first-principles simulations, I will show that the surface of common electrolyte solutions is stratified into two distinct water layers: one depleted of ions and the other enriched with them.

Next, I will present our recent investigation into water autodissociation [3]. We employ the modern theory of polarization to perform periodic ab initio molecular dynamics simulations of water under external electric fields. Our simulations reveal that the enhancement of water dissociation in these conditions is primarily driven by entropic effects rather than enthalpic ones, as is normally assumed. Finally, I will discuss how these findings may provide crucial insights into recent kinetic measurements of the hydrogen evolution reaction (HER) across various electrochemical systems [4].

[1] Y. Litman, J. Lan, Y Nagata, D. M. Wilkins, J. Phys. Chem. Lett. 14, 8175-8182 (2023) [2] Y. Litman, K-Y. Chiang, T. Seki, Y. Nagata, M. Bonn, Nat. Chem. (2024) 16, 644–650 (2024) [3] Y. Litman, A. Michaelides (in preparation) [4] J. M. Gisbert-González, C. G. Rodellar, J. Druce, E. Ortega, B. Roldan Cuenya, S. Z. Oener, J. Am. Chem. Soc (in press, DOI : 10.1021/jacs.4c18638)

Add to your calendar or Include in your list

Wed 26 Feb 14:30: Probing biomolecular phase separation through multiscale computer simulations.

School of Physical Sciences - Mon, 10/02/2025 - 09:57
Probing biomolecular phase separation through multiscale computer simulations.

Biomolecular condensates play crucial roles in cellular organisation, regulating diverse biological functions, as well as contributing to disease pathologies when phase separation is dysregulated. However, the physicochemical mechanisms by which they are formed and regulated are still not well understood, especially in the complex environment inside cells consisting of thousands of different components. Computer simulations have emerged as powerful tools to investigate phase transitions in these systems. In this talk, we will discuss how coarse-grained molecular-dynamics simulations at different resolutions can probe the molecular mechanisms governing biomolecular phase separation across different systems, as well as guide the design of proteins that can give rise to condensates with specific properties.

Add to your calendar or Include in your list

Thu 17 Apr 14:00: Title to be confirmed

School of Physical Sciences - Sun, 09/02/2025 - 12:05
Title to be confirmed

Abstract not available

Add to your calendar or Include in your list

Thu 13 Feb 17:00: Algebraising foundations of elliptic curves

School of Physical Sciences - Sat, 08/02/2025 - 19:25
Algebraising foundations of elliptic curves

Elliptic curves are one of the simplest non-trivial objects in algebraic geometry, which are pervasive in modern number theory, but also see applications in point counting algorithms and public key cryptography. Due to their geometric nature, formalising a working definition typically requires a lot of technical machinery, let alone any non-trivial results. Yet, the Lean community has managed to formalise two of the most fundamental theorems in the theory of elliptic curves, with scope for many more projects. In this talk, I will explain these theorems, and how we inadvertently discovered new proofs in our formalisation attempts. Elliptic curves are one of the simplest non-trivial objects in algebraic geometry, which are pervasive in modern number theory, but also see applications in point counting algorithms and public key cryptography. Due to their geometric nature, formalising a working definition typically requires a lot of technical machinery, let alone any non-trivial results. Yet, the Lean community has managed to formalise two of the most fundamental theorems in the theory of elliptic curves, with scope for many more projects. In this talk, I will explain these theorems, and how we inadvertently discovered new proofs in our formalisation attempts.

=== Hybrid talk ===

Join Zoom Meeting https://cam-ac-uk.zoom.us/j/87143365195?pwd=SELTNkOcfVrIE1IppYCsbooOVqenzI.1

Meeting ID: 871 4336 5195

Passcode: 541180

Add to your calendar or Include in your list

Wed 19 Feb 16:30: TBC

School of Physical Sciences - Sat, 08/02/2025 - 08:47
TBC

TBC

Add to your calendar or Include in your list

Mon 17 Feb 14:00: (Cancelled)

School of Physical Sciences - Fri, 07/02/2025 - 18:35
(Cancelled)

Abstract not available

Add to your calendar or Include in your list

Wed 19 Feb 16:15: Searching for new physics with ultracold molecules

School of Physical Sciences - Fri, 07/02/2025 - 14:41
Searching for new physics with ultracold molecules

In the Standard Model of particle of physics, the electron has a tiny permanent electric dipole moment (EDM). In most theories that extend the Standard Model, this EDM is predicted to be many orders of magnitude larger due to new CP-violating mechanisms. Thus, EDM measurements are searches for new CP-violating physics which is deeply connected to the puzzle of the matter-antimatter asymmetry of the Universe. The most precise measurements of the electron EDM all use molecules. The molecules are spin polarized, and the EDM determined by measuring the spin precession frequency in an applied electric field. The precession is due to the interaction of the EDM with an effective electric field which can be exceptionally large for heavy polar molecules. To reach high precision we need long spin precession times, which is only possible with neutral molecules if they are cooled to microkelvin temperatures. I will present our efforts to measure the electron EDM using laser-cooled YbF molecules, both in a beam and, in the future, trapped in an optical lattice.

Add to your calendar or Include in your list

Fri 14 Feb 18:00: Modelling for the Budding Mathematician

School of Physical Sciences - Fri, 07/02/2025 - 13:13
Modelling for the Budding Mathematician

Both pure and applied mathematics are about modelling: one approximates concepts, while the other is concerned with ‘the real world’. Either form must contend with the inherent difficulties of the enterprise.

This talk will discuss what ‘modelling’ entails and how the tools of category theory provide natural ways to tackle the associated challenges. This is intended as a gentle introduction to categorical thinking, and no prior background is assumed. If time permits, I will also sketch an elegant new approach to modelling complex systems based on the theory of double categories and how it is connected to assume-guarantee reasoning.

Add to your calendar or Include in your list

Tue 11 Feb 11:15: The capacity of Constant Torsion Emergent Gravity to resolve cosmological tensions

School of Physical Sciences - Fri, 07/02/2025 - 11:07
The capacity of Constant Torsion Emergent Gravity to resolve cosmological tensions

Despite the success of the ΛCDM model in explaining a wide range of cosmological phenomena, observational discrepancies such as the Hubble tension and curvature tension, as well as theoretical challenges such as the inability to unify General Relativity with other fundamental forces in particle physics, have prompted a reevaluation of our current cosmological model and an exploration of other theories of gravity. We investigate the constant torsion emergent gravity (CTEG) model, a specific case of Poincare Gauge Theory (PGT) of gravity. This model introduces two extra cosmological parameters in addition to the six parameters in ΛCDM, and can be treated as a standard extension to ΛCDM with modified dark energy. This modification is then implemented with CAMB and the constraints of the cosmological parameters of PGT is obtained through nested sampling using Polychord and Cobaya. Our results offer a comparative analysis against the ΛCDM model, looking at the possibility of CTEG to resolve key observational tensions.

Add to your calendar or Include in your list

Fri 14 Feb 13:00: Third law of black hole mechanics for supersymmetric black holes

School of Physical Sciences - Fri, 07/02/2025 - 10:41
Third law of black hole mechanics for supersymmetric black holes

The third law of black hole mechanics asserts that it is impossible for a non-extremal black hole to become extremal in finite time (in classical General Relativity). A proof of this law was claimed in the 1980s. However, counterexamples to this law were found recently: gravitational collapse of a massless charged scalar field can produce an exactly extremal Reissner-Nordstrom black hole in finite time, passing through an intermediate phase in which the solution is exactly Schwarzschild at the horizon. These examples involve matter with a large charge to mass ratio. What about theories, such as supersymmetric theories, with an upper bound on the charge to mass ratio of matter? In this case I have proved that one cannot form a supersymmetric black hole (such as extremal Reissner-Nordstrom) in finite time. Thus a third law holds for supersymmetric black holes. The proof involves ideas related to quasi-local energy. In this talk I shall review all of these developments.

Add to your calendar or Include in your list