Dear All,
The next speaker in our CWI seminar for Machine Learning and Uncertainty Quantification for Scientific Computing will be Christian Franzke from IBS Center for Climate Physics at the Pusan National University in South Korea. The talk will be on Wednesday the 23rd, at 10AM CEST. The topic will be the detection of causal relationships and physically meaningful patterns from the complex climate system, using reservoir computing and multi-resolution dynamic mode decomposition. The zoom link and abstract can be found below. As usual, feel free to share the zoom link with your colleagues.
Kind regards,
Wouter Edeling
23 Jun. 2021 10h00 CET: Christian Franzke (IBS Center for Climate Physics, Pusan National University in South Korea): Causality Detection and Multi-Scale Decomposition of the Climate System using Machine Learning
Detecting causal relationships and physically meaningful patterns from the complex climate system is an important but challenging problem. In my presentation I will show recent progress for both problems using Machine Learning approaches. First, I will show that Reservoir Computing is able to systematically identify causal relationships between variables. I will show evidence that Reservoir Computing is able to systematically identify the causal direction, coupling delay, and causal chain relations from time series. Reservoir Computing Causality has three advantages: (i) robustness to noisy time series; (ii) computational efficiency; and (iii) seamless causal inference from high-dimensional data. Second, I will demonstrate that Multi-Resolution Dynamic Mode Decomposition can systematically identify physically meaningful patterns in high-dimensional climate data. In particular, Multi-resolution Dynamic Mode Decomposition is able to extract the changing annual cycle.
Join Zoom Meeting
https://cwi-nl.zoom.us/j/81854110402?pwd=YTBvNU9qWHlBaVA2aURISGtKeitSUT09
Meeting ID: 818 5411 0402
Passcode: 599921
Dear all,
On Thursday June 10th, at 4PM CET, Hannah Christensen from the University of Oxford will give the next talk in the CWI seminar on Machine Learning and Uncertainty quantification for Scientific computing. She will talk about stochastic subgrid-scale parametrisation using GANs, applied to atmospheric models. The abstract and zoom link can be found below. Feel free to share the zoom link with anyone who you think would be interested.
Kind regards,
Wouter Edeling
10 Jun. 2021 16h00 CET: Hannah Christensen (Oxford): Machine Learning for Stochastic Parametrisation
Atmospheric models used for weather and climate prediction are traditionally formulated in a deterministic manner. In other words, given a particular state of the resolved scale variables, the most likely forcing from the sub-grid scale motion is estimated and used to predict the evolution of the large-scale flow. However, the lack of scale-separation in the atmosphere means that this approach is a large source of error in forecasts. Over the last decade an alternative paradigm has developed: the use of stochastic techniques to characterise uncertainty in small-scale processes. These techniques are now widely used across weather, seasonal forecasting, and climate timescales.
While there has been significant progress in emulating parametrisation schemes using machine learning, the focus has been entirely on deterministic parametrisations. In this presentation I will discuss data driven approaches for stochastic parametrisation. I will describe experiments which develop a stochastic parametrisation using the generative adversarial network (GAN) machine learning framework for a simple atmospheric model. I will conclude by discussing the potential for this approach in complex weather and climate prediction models.
Join Zoom Meeting
https://cwi-nl.zoom.us/j/86156254037?pwd=QnpVTlZqdnpnaDdvNm83TlM3MTFZUT09
Meeting ID: 861 5625 4037
Passcode: 364032
The Department of Applied Mathematics of the University of Twente has a vacancy for an Assistant or Associate Professor (f/m) in Optimization and Learning with particular focus on mathematical methods for Markov decision theory and reinforcement learning applied in healthcare optimization. The position is embedded in both Operations Research and the Center for Healthcare Operations Improvement and Research (CHOIR). For details, see
https://www.utwente.nl/en/organisation/careers/!/2021-381/assistantassociat…
--------------------------------------
Prof.dr. R.J. Boucherie
people.utwente.nl/r.j.boucherie<http://people.utwente.nl/r.j.boucherie>
Stochastic Operations Research
Department of Applied Mathematics
University of Twente
www.utwente.nl/ewi/sor<http://www.utwente.nl/ewi/sor>
CHOIR (Center for Healthcare Operations Improvement & Research)
www.utwente.nl/choir<http://www.utwente.nl/choir>
P.O. Box 217
NL-7500 AE Enschede
P: +31 53 489 3432
E-mail: r.j.boucherie(a)utwente.nl
https://people.utwente.nl/r.j.boucherie
Are you a bioinformatician passionate about medical research and looking
for an exciting and impactful new challenge? Do you have experience in
analysing *genome-wide omics* data and *high-throughput molecular or
genetic screens* for target identification and drug discovery?
We are currently seeking a talented *Senior Bioinformatician* to join our
Data Sciences & Quantitative Biology team in *Cambridge, UK* and apply
bioinformatics and computational solutions to help us evolve and extend our
capabilities in our drive to find the medicines of the future. In doing so,
you will be part of a multi-disciplinary team of bioinformaticians, data
scientists, image analysts and statisticians providing quantitative
insights to discover new therapies. Our group is committed to enhancing
AstraZeneca’s ability to enable effective target selection and hit
identification.
We welcome applications no later than *Saturday 5th June*. For more details
and to apply, use the following link:
https://careers.astrazeneca.com/job/cambridge/senior-bioinformatician-molec…
(apologies for multiple posting)
https://jobs.tue.nl/en/vacancy/phd-on-datadriven-optimization-for-sustainab…<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fjobs.tue.…>
The “Atlas leefbare stad” (Atlas livable city) is a digital twin application that processes logistical data for urban environments. It helps policymakers to gain insight into the current logistical situation, but it can also be used to find the optimal decision with respect to an objective such as fuel consumption or greenhouse gas emission. This makes it possible to try new optimal solutions in a realistic virtual environment, before applying them in practice. As a PhD candidate, you are expected to develop new machine learning driven optimization algorithms, using techniques such as reinforcement learning, online supervised learning, Bayesian optimization, etc., that can handle such a complex data-based environment. Different research goals such as incorporating expert knowledge in the data-driven models and designing objective functions and uncertainty measures that are easy to learn and to optimize, make this a challenging project and require a mix of applied and theoretical research.
Dear all,
This Friday (the 21st) at 4PM CET, John Harlim from Penn State will present his work on machine learning of missing dynamical systems at CWI. Here, the model error problem arising from missing dynamics is formulated as a supervised learning task using the Mori-Zwanzig formalism and Taken's embedding theorem. More info can be found in the abstract below. Feel free to share the zoom link if you think the talk may be of interest to your colleagues or students.
Kind regards,
Wouter Edeling
Join Zoom Meeting
https://cwi-nl.zoom.us/j/86326565885?pwd=THBiY3BRSUkvVzQ1UjM4Y1RTNGhOZz09
Meeting ID: 863 2656 5885
Passcode: 931873
13 May 2021 16h00: John Harlim (Penn state): Machine learning of missing dynamical systems
In the talk, I will discuss a general closure framework to compensate for the model error arising from missing dynamical systems. The proposed framework reformulates the model error problem into a supervised learning task to estimate a very high-dimensional closure model, deduced from the Mori-Zwanzig representation of a projected dynamical system with projection operator chosen based on Takens embedding theory. Besides theoretical convergence, this connection provides a systematic framework for closure modeling using available machine learning algorithms. I will demonstrate numerical results using a kernel-based linear estimator as well as neural network-based nonlinear estimators. If time permits, I will also discuss error bounds and mathematical conditions that allow for the estimated model to reproduce the underlying stationary statistics, such as one-point statistical moments and auto-correlation functions, in the context of learning Ito diffusions.
Dear all,
Do you know someone interested in Space-time Neural Nets, Memory Models & (Differential) Geometry, Groups, Manifolds? We are looking for an enthusiastic and driven researcher for a Ph.D. application deadline by the beginning of June. For more details on the position, please check the vacancy post: https://www.uva.nl/en/content/vacancies/2021/05/21-326-phd-position-space-t…
The position will be supervised by Associate Professor Dr. Efstratios Gavves at the University of Amsterdam, director of QUVA Lab, and director of POP-AART Lab. The position is fully funded by ERC Starting Grant and NWO VIDI. For any questions, you can send an email at egavves(a)uva.nl.
Please, share!
Best,
Stratis
Dear all,
Do you know someone interested in Space-time Neural Nets in the context of Generative/Stochastic/Bayesian models? We are looking for an enthusiastic and driven researcher for a Ph.D. application deadline by the beginning of June. For more details on the position, please check the vacancy post: https://www.uva.nl/en/content/vacancies/2021/05/21-325-phd-position-generat…
The position will be supervised by Associate Professor Dr. Efstratios Gavves at the University of Amsterdam, director of QUVA Lab, and director of POP-AART Lab. The position is fully funded by ERC Starting Grant and NWO VIDI. For any questions, you can send an email at egavves(a)uva.nl.
Please, share!
Best,
Stratis
Dear all,
Do you know someone interested in unsupervised deep representation learning for video? We are looking for an enthusiastic and driven researcher for a Ph.D. application deadline by the beginning of June. For more details on the position, please check the vacancy post: https://www.uva.nl/en/content/vacancies/2021/05/21-335-phd-position-unsuper…
The position will be supervised by Associate Professor Dr. Efstratios Gavves at the University of Amsterdam, director of QUVA Lab, and director of POP-AART Lab. The position is fully funded by ERC Starting Grant and NWO VIDI. For any questions, you can send an email at egavves(a)uva.nl.
Please, share!
Best,
Stratis