Dr. Mats Brorsson
|
|
||||
| Faculté ou Centre | Interdisciplinary Centre for Security, Reliability and Trust | ||||
| Department | SEDAN | ||||
| Adresse postale |
Université du Luxembourg 29, avenue JF Kennedy L-1855 Luxembourg |
||||
| Bureau sur le campus | JFK Building, E02-207 | ||||
|
|||||
| Téléphone | (+352) 46 66 44 9455 | ||||
Mats Brorsson received his PhD degree from the Lund University (Sweden), in 1995. He is a Professor of Computer Architecture at KTH Royal Institute of Technology in Sweden, from which he is on a leave of absence. Since 2015 he has worked as software engineer in industrial companies based in Luxembourg. His research interests are in programming models, run-time systems, operating systems and the architecture of parallel computer systems, in particular multi- and manycore systems. Mats joined the Service and Data Management in Distributed Systems research group, SEDAN, headed by Dr. Habil. Radu State.
Last updated on: jeudi 23 mars 2023
Highlights in reverse date order:
- Research scientist at University of Luxembourg, 2019-
- Optimisation scientist at Satalia (www.satalia.com), 2017-2020
- Senior software engineer at OLAmobile (www.olamobile.com), 2015-2017
- Professor of Computer Architecture at KTH (www.kth.se), 2000- (on leave of absence since 2015)
- Assistant/Associate professor at Lund University (www.lu.se), 1994-2000
- PhD in Computer systems engineering, Lund university, 1994
- PhD student at Lund University, 1985-1989, 1991-1994
- Systems engineer, Telesoft AB, 1989-1991
- MSc in Electrical Engineering, Lund university, 1985
Last updated on: 16 avr 2021
Current projects
| ACE5G | MAELSTROM | SCRiPT |
ACE5G — Accelerated Cloud Edge in 5G
This is a project in cooperation with Luxembourgish Telecom and Cloud provider Proximus (www.proximus.lu).
Cloud computing has enabled elastic provisioning of computer resources and revolutionised the industry leading to basically zero cost investment needed for infrastructure of new IT-companies. Most resources provisioned this way comes from data centres with a vast array of quite similar compute offerings. With modern distributed client-facing applications, however, there is a need to move the resources closer to the clients, to the Edge. While this is possible today, it is a manual time-consuming and error prone process.
The ACE5G project will automate this process by developing a set of complementary services. We will develop a technology to allow software components to be executed on heterogeneous hardware (including accelerators) without being modified. We will also develop scheduling technology that can take performance constraints into account and run the software at the compute resources that make most sense taking not only performance into account but also power consumption.
MAELSTROM
This is a EuroHPC collaborative project in collaboration with six other partners across Europe.
To develop Europe’s computer architecture of the future, MAELSTROM will co-design bespoke compute system
designs for optimal application performance and energy efficiency, a software framework to optimise usability and
training efficiency for machine learning at scale, and large-scale machine learning applications for the domain of
weather and climate science.
The MAELSTROM compute system designs will benchmark the applications across a range of computing systems
regarding energy consumption, time-to-solution, numerical precision and solution accuracy. Customised compute
systems will be designed that are optimised for application needs to strengthen Europe’s high-performance computing
portfolio and to pull recent hardware developments, driven by general machine learning applications, toward needs of
weather and climate applications.
The MAELSTROM software framework will enable scientists to apply and compare machine learning tools and
libraries efficiently across a wide range of computer systems. A user interface will link application developers with
compute system designers, and automated benchmarking and error detection of machine learning solutions will be
performed during the development phase. Tools will be published as open source.
The MAELSTROM machine learning applications will cover all important components of the workflow of weather
and climate predictions including the processing of observations, the assimilation of observations to generate
initial and reference conditions, model simulations, as well as post-processing of model data and the development
of forecast products. For each application, benchmark datasets with up to 10 terabytes of data will be published
online for training and machine learning tool-developments at the scale of the fastest supercomputers in the world.
MAELSTROM machine learning solutions will serve as blueprint for a wide range of machine learning applications
on supercomputers in the future.
Our part of the project is mostly in working on benchmarks and on performance models of future computer systems for ML-augmented weather & climate modelling applications.
SCRiPT — SME Credit Risk Platform
This is a project in cooperation with Yoba Smart Money (www.yoba.com) and funded by FNR.
The SCRiPT project aims to develop models and methodologies to assess the credit risk of small and medium sized enterprises (SMEs) using data analytics, thus allowing Yoba to make credit decisions quicker and more efficiently and thus increasing the overall availability of credit to the underserved SME sector in Luxembourg and other European markets where Yoba will operate.
The ultimate goal is to increase the use of data-based credit scoring generally amongst SME lenders and potentially to increase the availability and use of business registers & credit rating agencies for this purpose.
The technology behind this platform will be fully supported by data and machine learning models will be utilised to: i) extract information from unstructured alternative data, ii) provide credit scoring, iii) make model drift analysis, and iv) make anomaly detection. Significant effort will be put on privacy aspects and explainable models.
Read more about it here.
Last updated on: 21 avr 2021
2023
Performance Analysis and Benchmarking of a Temperature Downscaling Deep Learning Model;
in 31st Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, Naples, Italy 1-3 March 2023 (2023, March)
2022
Performance Modeling of Weather Forecast Machine Learning for Efficient HPC;
in International Conference on Distributed Computing Systems (ICDCS), Italy 10-13 July 2022 (2022, October 13)
A Cloud-Edge Continuum Experimental Methodology applied to a 5G Core Study; ;
in Transactions on Computational Science and Computational Intelligence (2022)
2021
2019
Know Your Enemies and Know Yourself in the Real-Time Bidding Function Optimisation; ; ; ; ; ;
in Proceedings of the 19th IEEE International Conference on Data Mining Workshops (ICDMW 2019) (2019)
Time Series Modeling of Market Price in Real-Time Bidding; ; ; ; ;
in 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (2019, April)
Regression-based prediction for task-based program performance; ; ;
in Journal of Circuits, Systems and Computers (2019), 28(04), 1950060
