Wed 01 May 14:30: Universality for bootstrap percolation
In this talk I will give an overview of the proof of the “Universality Conjecture” for general bootstrap percolation models. Roughly speaking, the conjecture states that every d-dimensional monotone cellular automaton is a member of one of d+1 universality classes, which are characterized by their behaviour on sparse random sets. More precisely, it states that if sites of the lattice Z^d are initially infected independently with probability p, then the expected infection time of the origin is either infinite, or is a tower of height r for some r \in {1,...,d}. I will also describe an uncomputability result regarding the exponent of p at the top of the tower.
Based on joint work with Paul Balister, Béla Bollobás and Paul Smith.
- Speaker: Rob Morris (IMPA)
- Wednesday 01 May 2024, 14:30-15:30
- Venue: MR12.
- Series: Combinatorics Seminar; organiser: ibl10.
Tue 07 May 13:10: TBC
Abstract not available
- Speaker: Stefanie Felsberger
- Tuesday 07 May 2024, 13:10-14:00
- Venue: 1 Newnham Terrace, Darwin College.
- Series: Darwin College Humanities and Social Sciences Seminars; organiser: Janet Gibson.
Tue 07 May 13:10: TBC
Abstract not available
- Speaker: Stefanie Felsberger
- Tuesday 07 May 2024, 13:10-14:00
- Venue: 1 Newnham Terrace, Darwin College.
- Series: Darwin College Humanities and Social Sciences Seminars; organiser: Janet Gibson.
Tailoring cobalt spinel oxide with site-specific single atom incorporation for high-performance electrocatalysis
DOI: 10.1039/D4EE00058G, Paper Open Access   This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.Taeghwan Hyeon, Kangjae Lee, Jaehyuk Shim, Hyunsoo Ji, Jungho Kim, Hyeon Seok Lee, Heejong Shin, Megalamane S. Bootharaju, Kug-Seung Lee, Wonjae Ko, Jaewoo Lee, Kang Kim, Seungwoo Yoo, Sungeun Heo, Jaeyune Ryu, Seoin Back, Byoung-Hoon Lee, Yung-Eun Sung
Universal incorporation of metals into cobalt spinel oxide (CSO) has emerged as a versatile and promising strategy to enhance catalytic performance. However, the uncontrolled reactivity of early transition metal and...
The content of this RSS Feed (c) The Royal Society of Chemistry
Tue 07 May 11:30: The role of methane for chemistry-climate interactions: rapid radiative adjustments and climate feedbacks Zoom link: https://us02web.zoom.us/j/89826306833?pwd=cnNHSG9OWHRjVngzMGVMc2F0NnA4dz09
Methane (CH4), the second most important greenhouse gas directly emitted by human activity, is removed from the atmosphere via chemical degradation. The chemical sink of CH4 is influenced by temperature and the chemical composition of the atmosphere. It is further an important source of water vapour in the stratosphere and affects the ozone concentration in the troposphere and the stratosphere via secondary feedbacks.
The talk will focus on the role of these chemistry-climate interactions in numerical simulations with the chemistry-climate model EMAC perturbed by either CO2 or CH4 increase. For both forcing agents, CO2 and CH4 , so called rapid radiative adjustments are assessed in simulations with prescribed sea surface temperatures, as well as climate feedbacks in respective simulations using an interactive oceanic mixed layer.
The simulation set-up uses CH4 emission fluxes instead of prescribed CH4 concentrations at the lower boundary so that changes of the chemical sink can feed back on the atmospheric CH4 concentration without constraints.
The results show a shortening of the CH4 lifetime and, accordingly, a reduction of the CH4 mixing ratios in a warming and moistening troposphere. This decrease in CH4 also affects the response of tropospheric ozone.
Furthermore, recently an additional radiation scheme was implemented into the EMAC model, which represents the direct radiative effect of CH4 better and also accounts for the absorption by CH4 in the solar shortwave spectrum. With the new radiation scheme the effective radiative forcing for the same perturbation of CH4 emissions is larger, and individual rapid radiative adjustments, e.g. of clouds, are changed.
Zoom link: https://us02web.zoom.us/j/89826306833?pwd=cnNHSG9OWHRjVngzMGVMc2F0NnA4dz09
- Speaker: Laura Stecher, Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre (Engl: German Aerospace Center, Institute of Atmospheric Physics)
- Tuesday 07 May 2024, 11:30-12:30
- Venue: Chemistry Dept, Unilever Lecture Theatre and Zoom.
- Series: Centre for Atmospheric Science seminars, Chemistry Dept.; organiser: .
Thu 02 May 13:00: Policy Portfolio for Banks: Deposit Insurance and Ex-post Liquidity Injection
Banking crises pose significant threats to our economy, leading to the implementation of policy measures such as deposit insurance and liquidity injection to strengthen financial stability and optimize resource allocation efficiency. This paper investigates the dynamic interplay between deposit insurance and liquidity injection. Facing uncertainty regarding bank health and depositor liquidity shocks, policymakers decide liquidity injection based on withdrawals. While higher deposit insurance coverage can mitigate panic runs, it may undermine the effectiveness of liquidity injections. We demonstrate that liquidity injection overshadows deposit insurance. Consequently, the optimal policy portfolio entails zero deposit insurance, enhancing resource allocation efficiency but leading to more panic runs.
- Speaker: Lin Shen (Insead)
- Thursday 02 May 2024, 13:00-14:00
- Venue: W2.02.
- Series: Cambridge Finance Workshop Series; organiser: Cerf Admin.
Wed 01 May 15:05: Exploring novel (bio)molecular spaces by design – a dialogue between representation and generation
The machine learning tool box has revolutionized our ability to design novel molecular entities (e.g. proteins) well beyond what the natural repertoire has explored. Despite the incredible advances, the de novo generation of functional molecules in biological concepts remains an incredible challenge.
In this talk I will present some of the efforts in our group in designing both proteins and small molecules. Particularly emphasizing different modalities of molecular representation and the interplay with generative ML to facilitate the exploration of unimaginably large spaces of possibilities. Importantly, many ML-based approaches for molecular design fall short in terms of generalization and sampling off the learned distribution, I will present some ideas on how representation can help to overcome some of these limitations.
Finally, I will present some of the approaches developed in our group to embed function into the designer proteins and how we are suing these components in cellular systems to control the output of these complex biological devices.
Link to join virtually: https://cam-ac-uk.zoom.us/j/81322468305
This talk is being recorded.
- Speaker: Professor Bruno Correia - Laboratory of Protein Design and Immunoengineering, EPFL Swiss Federal Institute of Technology, Lausanne.
- Wednesday 01 May 2024, 15:05-15:55
- Venue: Lecture Theatre 1, Computer Laboratory, William Gates Building.
- Series: Wednesday Seminars - Department of Computer Science and Technology ; organiser: Ben Karniely.
Fri 07 Jun 13:00: Title to be confirmed
Abstract not available
- Speaker: Frank Feng, Independent Researcher
- Friday 07 June 2024, 13:00-14:00
- Venue: GN06, William Gates Building. Zoom link: https://cl-cam-ac-uk.zoom.us/j/4361570789?pwd=Nkl2T3ZLaTZwRm05bzRTOUUxY3Q4QT09&from=addon .
- Series: Energy and Environment Group, Department of CST; organiser: lyr24.
Thu 25 Apr 14:00: Multiresolution Mesh Rendering Engine - Practicalities and Performance
[CESCG practice talk]
A multiresolution mesh is a structure that allows multiple levels of resolution of a mesh to be sampled in different regions. They are used to accelerate the construction of view-dependent Levels of Detail (LODs) for real-time rendering, generally for complex objects that may span large depths (e.g. terrain). Nanite, introduced in Unreal Engine 5, is an example of a full multiresolution pipeline. We describe our mesh-shader based multiresolution rendering engine in Vulkan, with two implementations to extract view dependent LODs. The first implementation is based on the approach established by Nanite. Our alternative implementation has no intermediate buffers at the cost of less fine-grained control over regions of the multires- olution we explore. We finally evaluate the two methods against each other and traditional LOD chains, emphasis ing practicality and performance.
Zoom link: https://cam-ac-uk.zoom.us/j/87633156881?pwd=ck9pR3YvdThSTHV4Ny8waXVQa3FYdz09
- Speaker: Maxwell Pettett, Univ. of Cambridge
- Thursday 25 April 2024, 14:00-15:00
- Venue: SS03 - William Gates Building.
- Series: Rainbow Group Seminars; organiser: Rafal Mantiuk.
Mon 29 Apr 14:00: STRICHARTZ ESTIMATES FOR THE 2D AND 3D MASSLESS DIRAC-COULOMB EQUATIONS
The massless Dirac equation with a Coulomb potential is interesting both from a physical and a mathematical point of view; it appears in some physical models, for instance the 2D equation is used to describe the dynamics of carbon atoms in a sheet of non-perfect graphene, and on the mathematical side the homogeneity of degree -1 of the potential seems to have a critical behavior, as |x| goes to infinity, since Strichartz estimates are known to hold for potentials that decay faster and there are examples of potentials decaying slower such that the corresponding flows do not disperse. In this talk I will present a recent result concerning Strichartz estimates for the solutions of the massless Dirac-Coulomb equation in 2 and 3 dimension with additional angular regularity. It extends the result on R3 of Cacciafesta-Séré-Zhang and provides completely new estimates on R2. As an application we will discuss a local well-posedness result for a nonlinear system.
- Speaker: Elena Danesi
- Monday 29 April 2024, 14:00-15:00
- Venue: MR13.
- Series: Partial Differential Equations seminar; organiser: Dr Greg Taujanskas.
Fri 24 May 13:00: Title to be confirmed
Abstract not available
- Speaker: Anaïs Berkes - Department of Computer Science and Technology, University of Cambridge
- Friday 24 May 2024, 13:00-14:00
- Venue: FW26, William Gates Building. Zoom link: https://cl-cam-ac-uk.zoom.us/j/4361570789?pwd=Nkl2T3ZLaTZwRm05bzRTOUUxY3Q4QT09&from=addon .
- Series: Energy and Environment Group, Department of CST; organiser: CRSD Team.
Thu 23 May 17:00: Cambridge RNA Club - May session IN PERSON
Abstract not available
- Speaker: Speaker to be confirmed
- Thursday 23 May 2024, 17:00-18:00
- Venue: Gurdon Institute Tea Room - RNA Club.
- Series: Cambridge RNA Club; organiser: .
Tue 07 May 14:00: Title to be confirmed
Abstract not available
- Speaker: Jose Gilberto Atondo Siu (University of Cambridge)
- Tuesday 07 May 2024, 14:00-15:00
- Venue: Zoom webinar - link to follow.
- Series: Computer Laboratory Security Seminar; organiser: Hridoy Sankar Dutta.
Mon 06 May 14:00: Leading-order term expansion for the Teukolsky equation on subextremal Kerr black holes
The study of wave propagation on black hole spacetimes has been an intense field of research in the past decades. This interest has been driven by the stability problem for black holes and by questions related to scattering theory. On Kerr black holes, the analysis of Maxwell’s equations and the equations of linearized gravity, can be simplified by introducing the Teukolsky equation, which offers the advantage of being scalar in nature. After explaining this reduction, I will present a result providing the large time leading-order term for initially localized and regular solutions of the Teukolsky equation, valid for the full subextremal range of black hole parameters and for all spins. I will explain how such a development follows naturally from the precise analysis of the resolvent operator on the real axis. Recent advances in microlocal analysis are crucially used to establish the existence and mapping properties of the resolvent.
- Speaker: Pascal Millet (Ecole Polytechnique)
- Monday 06 May 2024, 14:00-15:00
- Venue: MR13.
- Series: Partial Differential Equations seminar; organiser: Dr Greg Taujanskas.
Tue 14 May 14:30: TBA
TBA
- Speaker: Luis García (UCL)
- Tuesday 14 May 2024, 14:30-15:30
- Venue: MR13.
- Series: Number Theory Seminar; organiser: Jef Laga.
Tue 28 May 14:30: TBA
TBA
- Speaker: Beth Romano (King's College London)
- Tuesday 28 May 2024, 14:30-15:30
- Venue: MR13.
- Series: Number Theory Seminar; organiser: Jef Laga.
Thu 25 Apr 15:00: Machine Learning and Dynamical Systems Meet in Reproducing Kernel Hilbert Spaces with Insights from Algorithmic Information Theory
Since its inception in the 19th century, through the efforts of Poincaré and Lyapunov, the theory of dynamical systems has addressed the qualitative behavior of systems as understood from models. From this perspective, modeling dynamical processes in applications demands a detailed understanding of the processes to be analyzed. This understanding leads to a model, which approximates observed reality and is often expressed by a system of ordinary/partial, underdetermined (control), deterministic/stochastic differential or difference equations. While these models are very precise for many processes, for some of the most challenging applications of dynamical systems, such as climate dynamics, brain dynamics, biological systems, or financial markets, developing such models is notably difficult. On the other hand, the field of machine learning is concerned with algorithms designed to accomplish specific tasks, whose performance improves with more data input. Applications of machine learning methods include computer vision, stock market analysis, speech recognition, recommender systems, and sentiment analysis in social media. The machine learning approach is invaluable in settings where no explicit model is formulated, but measurement data are available. This is often the case in many systems of interest, and the development of data-driven technologies is increasingly important in many applications. The intersection of the fields of dynamical systems and machine learning is largely unexplored, and the objective of this talk is to show that working in reproducing kernel Hilbert spaces offers tools for a data-based theory of nonlinear dynamical systems.
In the first part of the talk, we introduce simple methods to learn surrogate models for complex systems. We present variants of the method of Kernel Flows as simple approaches for learning the kernel that appear in the emulators we use in our work. First, we will discuss the method of parametric and nonparametric kernel flows for learning chaotic dynamical systems. We’ll also explore learning dynamical systems from irregularly sampled time series and from partial observations. We will introduce the methods of Sparse Kernel Flows and Hausdorff-metric based Kernel Flows (HMKFs) and apply them to learn 132 chaotic dynamical systems. We draw parallels between Minimum Description Length (MDL) and Regularization in Machine Learning (RML), showcasing that the method of Sparse Kernel Flows offers a natural approach to kernel learning. By considering code lengths and complexities rooted in Algorithmic Information Theory (AIT), we demonstrate that data-adaptive kernel learning can be achieved through the MDL principle, bypassing the need for cross-validation as a statistical method. Finally, we extend the method of Kernel Mode Decomposition to design kernels in view of detecting critical transitions in some fast-slow random dynamical systems.
Then, we introduce a data-based approach to estimating key quantities which arise in the study of nonlinear autonomous, control, and random dynamical systems. Our approach hinges on the observation that much of the existing linear theory may be readily extended to nonlinear systems – with a reasonable expectation of success – once the nonlinear system has been mapped into a high or infinite dimensional Reproducing Kernel Hilbert Space. We develop computable, non-parametric estimators approximating controllability and observability energies for nonlinear systems. We apply this approach to the problem of model reduction of nonlinear control systems. It is also shown that the controllability energy estimator provides a key means for approximating the invariant measure of an ergodic, stochastically forced nonlinear system. Finally, we show how kernel methods can be used to approximate center manifolds, propose a data-based version of the center manifold theorem, and construct Lyapunov functions for nonlinear ODEs.
- Speaker: Boumediene Hamzi
- Thursday 25 April 2024, 15:00-16:00
- Venue: Centre for Mathematical Sciences, MR2.
- Series: Applied and Computational Analysis; organiser: Matthew Colbrook.
Thu 02 May 19:15: Monitoring Vaccine Effectiveness: can we trust results from parties with a vested interest?
Two observational methods are currently being used to monitor post-deployment vaccine effectiveness against infection: the obvious crude method comparing rate of testing positive for infection per head of vaccinated population with that rate per head of unvaccinated population; and the test-negative case control (TNCC) method. The two methods give very different results. Various parties’ preference for choice of method appears to broadly coincide with their vested interests in getting the result that method gives. We want to know whether either method is reliable.
We suggest how this question could be examined, and will share what conclusions we reach.
- Speaker: Roger Sewell
- Thursday 02 May 2024, 19:15-21:30
- Venue: MRC Cognition and Brain Sciences Unit, 15 Chaucer Road, Cambridge CB2 7EF.
- Series: Cambridge Statistics Discussion Group (CSDG); organiser: Peter Watson.
Thu 02 May 13:30: Overcoming challenges to sustainable heat using physics-informed machine learning
Heat pumps represent a sustainable heating technology that will play a crucial role in achieving Net Zero and decarbonisation goals in the UK. Increasing data availability from Building Management Systems and the rise of physics-based learning algorithms offer a solution to the problem of assessing the energy efficiency and maintenance requirements of an aging national heat pump fleet.
- Speaker: Javier Sandoval
- Thursday 02 May 2024, 13:30-14:00
- Venue: Richard King room, Darwin College.
- Series: Darwin College Science Seminars; organiser: Dr Sandra Petrus-Reurer.
Wed 20 Nov 14:30: Title to be confirmed
Abstract not available
- Speaker: Dr Laura Filion, University of Utrecht
- Wednesday 20 November 2024, 14:30-15:30
- Venue: Unilever Lecture Theatre, Yusuf Hamied Department of Chemistry.
- Series: Theory - Chemistry Research Interest Group; organiser: Lisa Masters.