Fri 02 Feb 15:00: The intersection of design, engineering and craft - what we learn by collaboration
Structural engineering spans a wide spectrum of practice, and the work at Cake Industries is very much at the “applied” end, where theory becomes real. Based on several key projects this talk will endeavour to look at how working right at the intersections of design, engineering and making can lead to developments in what is possible.
- Speaker: David Knight, Cake Industries
- Friday 02 February 2024, 15:00-16:00
- Venue: CivEng Seminar Room (1-33) (Civil Engineering Building).
- Series: Engineering Department Structures Research Seminars; organiser: Callum White.
Mon 12 Feb 13:00: The Influence of Lateral Spreading upon Solitary Wave Formation by Internal Tides
Abstract not available
- Speaker: Bruce Sutherland, University of Alberta
- Monday 12 February 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Wed 01 May 14:30: Sequence-based determinants of aggregation within protein condensates
Complex cellular landscapes of proteins include the dense, liquid-like droplet state and the solid-like amyloid state, in addition to the native state. The amyloid state, which is often pathological, can be formed through the deposition pathway from the native state and through the condensation pathway from the droplet state. I present a uniform framework to describe both pathways and identify mutations biasing towards these aggregation mechanisms. The droplet landscape model is a sequence-based, generic approach that simultaneously estimates the probability of droplet formation and the likelihood of state conversion. The method exploits that the interactions driving the droplet state sample disordered binding modes, whereas those governing the amyloid state sample ordered binding modes, which can simultaneously be estimated from sequence without information on the interaction partners. In addition, we predict the multiplicity of binding modes, that a given protein region can sample under different cellular conditions. I will demonstrate the application of the droplet landscape approach to both pathological and functional aggregates, in particular predicting mutations associated with amyotrophic lateral sclerosis and those facilitating muscle lineage development.
References M. Vendruscolo, M Fuxreiter (2022) Protein Condensation Diseases: Therapeutic Opportunities. Nat Commun 13, 5500, doi: 10.1038/s41467-022-32940-7 Hatos A, Tosatto SCE , Vendruscolo M, Fuxreiter M. (2022) FuzDrop on AlphaFold: visualizing the sequence-dependent propensity of liquid-liquid phase separation and aggregation of proteins. Nucleic Acids Res. 50(W1), W337 -44 Gönczi M., Teixeira JMC , Barrera-Vilarmau S., Mediani L. , Antoniani F. , Nagy TM, Fehér K., Ráduly Z., Ambrus V., Tőzsér J., Barta E., Kövér KE., Csernoch L., Carra S. , Fuxreiter M. (2023) Alternatively spliced exon regulates context-dependent MEF2D higher-order assembly during myogenesis Nature Communications 14, 1329. Horvath A, Vendruscolo M, Fuxreiter M. (2022) Sequence-based Prediction of the Cellular Toxicity Associated with Amyloid Aggregation within Protein Condensates Biochemistry 61, 2461-2469.
- Speaker: Professor Monika Fuxreiter, University of Padova
- Wednesday 01 May 2024, 14:30-15:30
- Venue: Unilever Lecture Theatre, Yusuf Hamied Department of Chemistry.
- Series: Theory - Chemistry Research Interest Group; organiser: Lisa Masters.
Mon 29 Jan 13:30: Bigger Picture Talks @ CEB - Pre-Procedural Planning of Transcatheter Heart Valve Interventions with Clinical Imaging and in silico Modelling
The extensive use of transcatheter devices has resulted in a paradigm shift in the clinical workflow for structural heart intervention. Pre-procedural planning, procedural guidance and follow-up care are critical to the success of these interventions. With a growing number of transcatheter heart valves available and a host of adverse outcomes to avoid, current decision-making processes are guided by clinical imaging, with 3D printing and image-based virtual simulation being recently adopted. However, these modalities cannot adequately capture or predict the dynamic interaction between implanted transcatheter devices and native anatomy. Thus, in silico models can be integrated as an additional tool to support the clinical decision-making processes. This lecture will provide an overview of recent advances in patient-specific in silico modelling of THV interventions and highlight the potential for clinical imaging with simulations to aid clinicians in patient selection and planning for transcatheter replacements of the aortic, mitral, and tricuspid valves.
- Speaker: Shelly Singh-Gryzbon
- Monday 29 January 2024, 13:30-14:30
- Venue: Lecture Theatre 1, Department of Chemical Engineering and Biotechnology, West Cambridge Site.
- Series: Chemical Engineering and Biotechnology; organiser: Alex Wilby.
Fri 01 Mar 16:00: High Performance Computing (HPC): what is it? how can I use it?
Computation is an important aspect of most scientific research these days. I will describe what is generally meant by “high performance” computing, and highlight different “workloads”, such as simulation, data processing and machine learning. I will also talk about locally available systems, national and international infrastructure. Finally, I will mention how to use the machines, whether that is with your own code, or using a well-known package, such as ASPECT or SPECFEM3D . I am also keen to hear what people are interested in doing, and happy to help them get started.
- Speaker: Chris Richardson
- Friday 01 March 2024, 16:00-17:00
- Venue: Tea Room, Old House.
- Series: Bullard Laboratories Tea Time Talks; organiser: David Al-Attar.
Thu 22 Feb 17:00: Title to be confirmed
Abstract not available
- Speaker: Dr Wenda Li (University of Edinburgh), Artem Khovanov (University of Cambridge), Michael Nedzelsky (Diffblue Ltd)
- Thursday 22 February 2024, 17:00-18:00
- Venue: MR14 Centre for Mathematical Sciences.
- Series: Formalisation of mathematics with interactive theorem provers ; organiser: Anand Rao Tadipatri.
Wed 07 Feb 15:05: The Compilation Game: Unifying AI, Hardware Design, Quantum, Climate Modelling, and Verification
Despite immense innovation pressure in the industry, we are held back by the slow evolution of our CPU -focused stand-alone compilation toolchains. Building a new domain-specific compiler, writing a new verification tool, optimizing an application, designing a microprocessor, or verifying some of its components: each of these tasks takes years. While the underlying problems are inherently complex, our inability to broadly exploit synergies across communities slows us down even more. Deep learning, battery electric vehicles, and rocket launches have seen orders-of-magnitude improvements over the last ten years, but compiler development is still slow. We must radically change the compiler development process: break it into pieces, scale the communities involved, use verification to enable scalability, and aggressively pursue automation across the stack. Open source can serve as a platform for this change, and our research in the context of the LLVM /MLIR community takes the first steps in this direction. I show how the number of compiler abstractions exploded recently, offer insight into the new hardware design stack CIRCT , and share our most recent efforts in high-performance computing and interactive theorem proving. Together, we will explore how these seemingly unrelated topics seed “the Compilation Game.”
Link to join virtually: https://cam-ac-uk.zoom.us/j/81322468305
This talk is being recorded.
- Speaker: Dr Tobias Grosser - Department of Computer Science and Technology, University of Cambridge
- Wednesday 07 February 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.
Mon 04 Mar 13:00: TBA
Abstract not available
- Speaker: Molly Menzel, NASA GIS
- Monday 04 March 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Mon 26 Feb 13:00: TBA
Abstract not available
- Speaker: Chiara Calascibetta
- Monday 26 February 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Mon 19 Feb 13:00: TBA
Abstract not available
- Speaker: Luke Kearney, University of Oxford
- Monday 19 February 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Mon 05 Feb 13:00: TBA
Abstract not available
- Speaker: Costanza Rodda, Imperial College
- Monday 05 February 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Mon 12 Feb 13:00: TBA
Abstract not available
- Speaker: Bruce Sutherland, University of Alberta
- Monday 12 February 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Wed 08 May 14:30: Title to be confirmed
Abstract not available
- Speaker: Professor Christian Jorgensen,
- Wednesday 08 May 2024, 14:30-15:30
- Venue: Unilever Lecture Theatre, Yusuf Hamied Department of Chemistry.
- Series: Theory - Chemistry Research Interest Group; organiser: Lisa Masters.
Thu 01 Feb 11:30: Transport and Settling of Buoyant Microplastics in Turbidity Currents
Although tens of millions of tons of plastic waste are released into the ocean each year, less than 300 kilotons remain on or near the ocean surface. This is particularly puzzling because more than half of plastics that are produced are buoyant in sea water. One mechanism that can result in buoyant plastic settling is the process of biofouling in which microbes and other organic material can accumulate on the plastics rendering them more dense. Less studied is the accumulation of inorganic material on the plastics. For example, clay has recently been shown to attach to plastics, particularly in the presence of surfactants. Here we report on laboratory experiments showing that plastic particles which are less dense than fresh water can settle due to the accumulation of glass spheres (“sand”) on their surface. This process is shown to occur dynamically as sand and plastic particles mix turbulently during the impulsive release of a turbidity current, which can carry some of the plastic particles to depth along with the settling sand. [This work reports on experiments performed by Woods Hole Oceanographic Institute (WHOI) Geophysical Fluid Dynamics (GFD) Fellow Quentin Kriaa during the WHOI GFD Summer Program 2023, co-supervised by Claudia Cenedese and Jim McElwaine.]
- Speaker: Professor Bruce Sutherland, Uni of Alberta
- Thursday 01 February 2024, 11:30-12:30
- Venue: Open Plan Area, Institute for Energy and Environmental Flows, Madingley Rise CB3 0EZ.
- Series: Institute for Energy and Environmental Flows (IEEF); organiser: Catherine Pearson.
Thu 15 Feb 15:00: Adaptive Intrusive Methods for Forward UQ in PDEs
In this talk we discuss a so-called intrusive approach for the forward propagation of uncertainty in PDEs with uncertain coefficients. Specifically, we focus on stochastic Galerkin finite element methods (SGFEMs). Multilevel variants of such methods provide polynomial-based surrogates with spatial coefficients that reside in potentially different finite element spaces. For elliptic PDEs with diffusion coefficients represented as affine functions of countably infinitely many parameters, well established theoretical results state that such methods can achieve rates of convergence independent of the number of input parameters, thereby breaking the curse of dimensionality. Moreover, for nice enough test problems, it is even possible to prove convergence rates afforded to the chosen finite element method for the associated deterministic PDE . However, achieving these rates in practice using automated computational algorithms remains highly challenging, and non-intrusive multilevel sampling methods are often preferred for their ease of use. We discuss an adaptive framework that is driven by a classical hierarchical a posteriori error estimation strategy — modified for the more challenging parametric PDE setting — and present numerical results.
- Speaker: Catherine Powell (University of Manchester)
- Thursday 15 February 2024, 15:00-16:00
- Venue: Centre for Mathematical Sciences, MR14.
- Series: Applied and Computational Analysis; organiser: Nicolas Boulle.
Fri 01 Mar 15:15: Epimorphisms and Acyclic Types in Univalent Mathematics
We characterize the epimorphisms in homotopy type theory (HoTT) as the fiberwise acyclic maps and develop a type-theoretic treatment of acyclic maps and types in the context of synthetic homotopy theory. We present examples and applications in group theory, such as the acyclicity of the Higman group, through the identification of groups with 0-connected, pointed 1-types. Many of our results are formalized as part of the agda-unimath library.
j.w.w. Ulrik Buchholtz and Egbert Rijke
- Speaker: Tom de Jong, University of Nottingham
- Friday 01 March 2024, 15:15-16:15
- Venue: SS03, Computer Laboratory.
- Series: Logic and Semantics Seminar (Computer Laboratory); organiser: Jon Sterling.
Fri 16 Feb 12:00: Scaling Multilingual Generation for Low-Resource Languages
The availability of large, high-quality datasets has been one of the main drivers of recent progress in generation tasks like summarization, QA. Such annotated datasets however are difficult and costly to collect, and rarely exist in languages other than English, rendering the technology inaccessible to underrepresented languages. An alternative to building large monolingual training datasets is to leverage pre-trained language models (PLMs). The talk will first discuss an approach, QAmeleon, that tunes a PLM using parameter efficient fine-tuning methods (PEFT) to synthesize QA data with only five examples per language. Using this data during training delivers accuracy superior to translation-based baselines, bridges nearly 60% of the gap between an English-only baseline and a fully supervised upper bound trained on almost 50,000 hand labeled examples. Next, the talk will discuss cross-lingual transfer approach for a much stricter zero-shot setting to enable generation in unseen languages. Our method composes language and task PEFT modules via element-wise arithmetic operations to leverage unlabeled data and labeled data in other languages. The talk further studied the consistency for cross-lingual generation tasks i.e. the output is in a language different that the source. Here we propose MuPlan which uses intermediate plans resulting in more faithful generation in both fine-tuning and zero-shot setups.
Bio:
Priyanka Agrawal is a Senior Research Scientist at the Google Deepmind in London, formally part of Google Brain, and is focused on building responsible Generative AI models and scaling them to underrepresented languages. Prior to that she was a Senior Researcher and Lead at http://Booking.com and IBM Research Labs, where she was driving work in cross-domain transfer and representation learning. She is an alumni from Computer Science Department at Indian Institute of Science. Her work is published at top tier ML and NLP conferences like NeurIPS, ACL and she holds 25+ US Patents. Priyanka also serves as Area Chair and PC member at these conferences and has been an invited panelist and speaker at various ML/NLP and diversity forums.
- Speaker: Priyanka Agrawal, Google Deepmind
- Friday 16 February 2024, 12:00-13:00
- Venue: Computer Lab, SS03.
- Series: NLIP Seminar Series; organiser: Eric Chamoun.
Mon 29 Jan 13:00: Negative effective viscosity in strongly stratified turbulence
Abstract not available
- Speaker: Qi Zhou, University of Calgary
- Monday 29 January 2024, 13:00-14:00
- Venue: MR5, CMS.
- Series: Geophysical and Environmental Processes; organiser: Prof. John R. Taylor.
Thu 01 Feb 15:00: What happens when you chop an equation?
This talk will discuss a tricky business: truncating a differential equation to produce finite solutions. A truncation scheme is often built directly into the steps needed to create a numerical system. E.g., finite differences replace exact differential operators with more manageable shadows, sweeping the exact approach off the stage.
In contrast, this talk will discuss the “tau method” which adds an explicit parameterised perturbation to an original equation. By design, the correction calls into existence an exact (finite polynomial) solution to the updated analytic system. The hope is that the correction comes out minuscule after comparing it with a hypothetical exact solution. The tau method has worked splendidly in practice, starting with Lanczos’s original 1938 paper outlining the philosophy. However, why the scheme works so well (and when it fails) remains comparably obscure. While addressing the theory behind the Tau method, this talk will answer at least one conceptual question: Where does an infinite amount of spectrum go when transitioning from a continuous differential equation to an exact finite matrix representation?
- Speaker: Geoff Vasil (University of Edinburgh)
- Thursday 01 February 2024, 15:00-16:00
- Venue: Centre for Mathematical Sciences, MR14.
- Series: Applied and Computational Analysis; organiser: Nicolas Boulle.