Active Projects

Numerical Algorithms, Frameworks, and Scalable Technologies for Extreme-Scale Computing

Computing has been disruptive to all scientific domains that increasingly rely on computational models and data to discover new knowledge and form decisions. With the explosion of Big Data, we are now faced with the ever-increasing size, variability, and uncertainty of the datasets. Some of the most challenging problems in data-driven science involve understanding the interactions between millions or even thousands of millions of variables. The vast quantity, veracity, velocity, and variety of data are challenging classical high-performance numerical methods and software for extreme-scale computing. Progress in research in scientific computing algorithms and software has been tightly linked to progress in microprocessor technology and high-performance programming technology. We are now in the process of embarking on the extreme-scale computing era which will revolutionize the architectural stack in a holistic fashion. It will also require research on optimized mathematical software libraries according to the device characteristics with novel numerical algorithms and data science applications that exploit them. How can we reconcile sustainable advances in sparse linear algebra and nonlinear optimization for new applications domains in data analytics while at the same time prepare for the anticipated sea-change looming in a twenty-year hardware horizon as well? We seek answers to these questions through computational methods that resolve fundamental challenges imposed by large-scale analytics, deep analysis, and precise predictions by advancing and preparing the foundation for the next generation of sparsified numerical methods. Our algorithms rely on the innovative coupling of sparsified numerical linear algebra and nonlinear optimization methods for data-intensive applications. The inherently deterministic character of these methods, when coupled with high communication demands, requires the development of robust approximation methods under the condition of extreme-scale computational science. This includes scientific libraries providing high-quality, reusable software components for constructing applications, as well as improved robustness ad portability. These developments will be driven by research on mathematical software, extreme-scale computing and an effort to push these developments toward applications. The focus on the computation of functions of matrix inverse entries presents a new dimension of numerical methods, since it goes beyond the classical requirements in solving linear systems or eigenvalue problems and has not yet been addressed in most of the research projects on massively parallel architectures. It is expected that the techniques developed by this will prove important in many of the other data-driven applications and will also provide basic tools for most of the applications for high performance computing (HPC) science and engineering problems. Novel, scalable software engineered to facilitate broader public use will be made available to the research and industrial community. Our numerical algorithms and mathematical software libraries are capable of leveraging emerging hardware paradigms and are applicable to a wide variety of existing applications such as finance, biology, health sciences, and many more. In particular, we will shed light on applications on nanoelectronic device simulation, and high-dimensional partial correlation estimations in genomics applications.

People

Olaf Schenk (Responsible)
Research Group
Advanced Computing Laboratory
Start Date
01.10.2022
End Date
30.09.2025
Duration
3 years
Funding Sources
SNSF
Status
Approved
Category
Swiss National Science Foundation / Lead Agency / Bilateral agreement with Germany

EUMaster4HPC - HPC European Consortium Leading Education Activities

Advancing education and training in High Performance Computing (HPC) and its applicability to HPDA and AI is essential for strengthening the world-class European HPC ecosystem. It is of primary importance to ensure the digital transformation and the sustainability of high-priority economic sectors. Missing educated and skilled professionals in HPC/HPDA/AI could prevent Europe from creating socio-economic value with HPC. The HPC European Consortium Leading Education Activities aims to develop a new and innovative European Master programme focusing on high performance solutions to address these issues. The master programme aims at catalysing various aspects of the HPC ecosystem and its applications into different scientific and industrial domains. EUMaster4HPC brings together major players in HPC education in Europe and mobilizes them to unify existing programs into a common European curriculum. It leverages experience from various European countries and HPC communities to generate European added value beyond the potential of any single university. EUMaster4HPC emphasizes collaboration across Europe with innovative teaching paradigms including co-teaching and the cooperative development of new content relying on the best specialists in HPC education in Europe. Employers, researchers, HPC specialists, supercomputing centres, CoEs and technology providers will constitute a workforce towardEuroHPC projects this master in HPC pilot programme. This pilot will provide a base for further national and pan- European educational programmes in HPC all over Europe and our lessons learned and the material development will accelerate the uptake of HPC in academia and industry.The creation of a European network of HPC specialists will catalyse transfers and mutual support between students, teachers and industrial experts. A particular focus on mobility of students and teachers will enable students to rapidly gain experience through internships and exposure to European supercomputing centres.

People

Olaf Schenk (Responsible)
Patrick Thomas Eugster, Ernst Wit (Co-responsibles)
External Partner
University of Luxembourg
Start Date
01.09.2022
End Date
31.08.2025
Duration
3 years
Funding Sources
European Commission
Status
Approved
Category
European and International Programmes / Horizon 2020 / EuroHPC

MaxEnt-Fin - Computational maximum entropy approach to high-dimensional modeling and analysis in finance

People

Illia Horenko (Responsibile)
Patrick Gagliardini (co-responsible)
Group
High-Dimensional Data Analysis Laboratory
Start Date
01.01.2022
End Date
31.12.2025
Duration
4 Years
Funding Sources
SNSF
Status
Active
Category
Swiss National Science Foundation / Project Funding / Division II - Mathematics, Natural and Engineering Sciences
The central goal of this project is to access the quality and performance of the existent and emerging machine learning (ML) and artificial intelligence (AI) approaches with respect to their ability to describe, to explain and to predict the neuronal behaviour on the basis of lab-data from mouse experiments. More common ML and AI approaches (hidden Markov models, shallow and reinforced learning, machine learning) will be compared to the very recently-developed Scalable Probabilistic Approximation approaches (Gerber et al., Sci. Adv. 2020) and to the entropy-driven approaches. Results of these comparison will aim at identifying the simplest possible (but not simpler then necessary) models that provide the most adequate lab-data descriptions. Identification of such models will enhance our understanding of emergence in the neuronal activity and provide a guidance for further experiments.

People

Horenko I. (Responsible)
Group
High-Dimensional Data Analysis Laboratory (Prof. Horenko)
Start Date
01.01.2021
End Date
31.12.2023
Duration
24 Months
Funding Sources
Carl Zeiss Foundation
Status
Active
Category
Carl Zeiss Foundation

People

E. C. Wit (Responsible)
Lomi A.(Co-Responsible)
Group
Statistical Computing Laboratory (Prof. Wit)
Start Date
01.09.2020
End Date
31.08.2024
Duration
48 Months
Funding Sources
SNSF
Status
Active
Category
Swiss National Science Foundation / Project Funding / Division I - Humanities and Social Sciences
Efficient energy trading relies on high-fidelity price prediction systems with short response times. This project will advance state-of-the-art computational tools for the market pricing mechanism for meeting real-time responses in smart power grid analysis.

People

Schenk O. (Responsible)
Group
Advanced Computing Laboratory (Prof. Schenk)
Start Date
01.08.2020
End Date
31.07.2023
Duration
36 Months
Funding Sources
Innosuisse
Status
Active
Category
Innosuisse / Innovation projects / Industrial partner: DXT Commodities
Progress in modern computing platforms and storage systems, electronic devices, and monitoring equipment has resulted in an exponential growth of the volume of data produced in several areas of science and engineering. These areas comprise of environmental sciences, biology= and medicine, satellite imaging, geospatial data, climate data, and transaction data among many others. Data processing commonly employs sophisticated statistical methods aiming to enrich the mechanisms governing the underlying physical processes and improve statistical models. Statistical analysis of such models traditionally has been carried out using Markov chain Monte Carlo methods (MCMC) used to represent complex dependency structures in data. MCMC methods provide a relatively simple approach to compute large hierarchical models requiring integration over several thousands of unknown parameters. Although MCMC methods are asymptotically exact they have slow convergence, do not scale well, and may fail for some complex models. It was soon realized that MCMC will not be able to meet modern and future big data challenges. In particular, we need to focus on extending the RINLA software ecosystem by advancing direct sparse linear solvers designed for Bayesian inference statistical computing. The sparse matrix algorithms and software implementations will be done in a codesign with data science applications in mind. By combining accelerated matrix algorithms and Bayesian inference at large scale, we plan to develop an algorithmic tool serving as part of a virtual laboratory for spatial and spatio-temporal models. This will pave the way for the next generation of data science applications based on INLA in ways not possible before. Our goal is to make our software ecosystem as productive and sustainable as possible by simultaneously focusing on algorithmic improvements to increase quality and speed, while at the same time evaluating potential benefits in various data science applications. This research project will therefore focus on solving all these fundamental challenges imposed by large-scale analytics, deep analysis and precise predictions by advancing and preparing the foundation for the next generation of RINLA.

People

Schenk O. (Responsible)
Haavard Rue (Co-Responsible)
Group
Advanced Computing Laboratory (Prof. Schenk)
Start Date
01.07.2020
End Date
31.06.2023
Duration
36 Months
Funding Sources
King Abdullah University of Science and Technology’s Competitive Research Grant Program
Status
Active
Category
King Abdullah University of Science and Technology’s Competitive Research Grant Program

Can Economic Policy Mitigate Climate-Change?

In this research project we plan to analyze possible economic responses to climate change in a heterogeneous-agents, multi-region, stochastic general equilibrium model. Climate change, as well as carbon taxation, will have drastically different effects on aggregate production and consumption across different regions. Moreover, a lack of international risk sharing as well as high costs to migration imply that the predicted global warming can have much larger adverse effects than it would appear from a single-region model. An obvious policy response to global warming is a carbon tax which will naturally hurt some individuals and help others. We plan to compute the optimal carbon tax through time, as well as region- and cohort-specific side payments needed to make carbon taxation a global, that is to say, an inter-temporal, and inter-regional win-win. To do so in an accurate quantitative fashion, we will need to i) develop large-scale, multi-region dynamic stochastic economic models with overlapping generations that incorporate state-of-the-art climate physics, and ii) develop high-performance computing codes that are capable of solving such models on a human time scale. An essential aspect of the research project is to develop economic models that can help us to understand how researchers and society can tackle the significant uncertainties associated with climate change. In this context, we also plan to address the question of how new financial assets or new forms of social insurance systems can help to share climate risks and mitigate climate uncertainties. Our project lies at the intersection of economics, climate science, and computational science. The main questions we ask are economic questions. However, to model climate change appropriately, in particular in order to quantify regional differences and uncertainties associated with climate change we need to engage and interact heavily with the climate modeling community. To compute the effects of taxes and climate risks on individuals’ welfares we plan to develop a modular code framework, with one module to model the evolution of climate, one module that links changes in climate to economic damages, and one module that solves for prices and quantities in the economy. For this, we need to interact heavily with the computational science community.

People

Olaf Schenk (Responsible)
Felix Kübler (Co-responsible)
External Partner
University of Zurich (principal beneficiary)
Start Date
01.12.2019
End Date
30.11.2023
Duration
4 years
Funding Sources
SNSF
Status
Active
Category
Swiss National Science Foundation / Sinergia
At the turn of the 21st century scientists have come to realise that a major ingredient in many modern economic, epidemiological, ecological and biological questions is to understand the network structure of the entities they study; for example, interbank lending is crucial for oiling the global economy and modern transport networks are facilitating the spread of infectious diseases. Unfortunately, even in the era of big data, computational bottle-necks have meant that only the simplest analyses have been applied to these large datasets, whereas methodological bottle-necks prevented an integrative view of complex phenomena. In short, inferring and analyzing complex networks have proven extremely difficult. Rather than simplifying the methodology prior to seeing the data, modern techniques from high-dimensional inference allow the data to select the appropriate level of complexity. The aim of this project is to integrate these techniques to the field of network analysis.

People

E. C. Wit (Responsible)
Group
Statistical Computing Laboratory (Prof. Wit)
Start Date
01.12.2019
End Date
30.11.2023
Duration
48 Months
Funding Sources
SNSF
Status
Active
Category
Swiss National Science Foundation / Project Funding / Division II - Mathematics, Natural and Engineering Sciences
Graph partitioning is a technique which is widely used in many fields of computer science and engineering. The goal is to partition the vertices of a graph into a certain number of disjoint sets of approximately the same size, so that a cut metric is minimized. Due to the NP-hardness of the problem and its practical importance, many different heuristics (spectral, combinatorial, evolutionist, etc.) have been developed to provide an approximate result in acceptable computational time. However, the advent of increasingly larger instances in emerging applications such as social networks or power networks renders graph partitioning and related graph algorithms such as clustering or community detection more and more important, multifaceted, and challenging. Spectral graph partitioning in the 2-norm using the eigenvectors of the graph Laplacian matrix was pioneered by Fiedler. However, this approach is considered infeasible for large-scale graphs, due to the prohibitively expensive computation of the Fiedler eigenvector, a fact that prompted the development of modern multilevel graph partitioning methods. Additionally, it has been proven that the quality of the partitions obtained by the spectral approach can be improved by considering the equivalent eigenvalue problem in the $p$-norm. In this project we plan to extend the aforementioned p-Laplacian partitioning method by developing two simultaneous research directions aiming at high quality graph partitioning on modern high performance parallel architectures. The first one will utilize the structural information encoded in the third eigenvector of the graph Laplacian matrix, corresponding to the third smallest eigenvalue. Incorporating more information present in the spectra has proven to improve the traditional spectral bisection algorithm, and such an approach is expected to improve the quality of the p-Laplacian partitioning scheme as well. Moreover, we intend to study the benefits of replacing the traditional gradient descent method currently utilized, with an optimization technique tailored for nonconvex scenarios.

People

Schenk O. (Responsible)
Group
Advanced Computing Laboratory (Prof. Schenk)
Start Date
01.04.2019
End Date
31.03.2023
Duration
48 Months
Funding Sources
SNSF
Status
Active
Category
Swiss National Science Foundation / Project Funding / Division II - Mathematics, Natural and Engineering Sciences