First Finnish LUMI Projects Chosen

July 20, 2022 — The first Finnish research projects to utilize LUMI have been chosen. The projects will research, e.g., dark matter, cancer research, digital twins of the Earth, translation and language models, and fusion energy research.

The resources were granted in the First Finnish LUMI Extreme Scale Call, which shared resources from Finland’s country share of the LUMI resources. The projects will start to run on LUMI, exploiting the system’s GPU capacity, in October.

The accepted projects are:

AXCESS: Axion dark matter from Exascale simulations, PI Kari Rummukainen, University of Helsinki, co-PI Mark Hindmarsh.
This project will work on simulations related to axions. A quarter of the matter in the Universe consists of an unknown particle interacting very weakly with ordinary matter. A leading candidate for this dark matter is the axion, which was proposed to solve a different fundamental problem in physics relating to the force between quarks inside protons and neutrons.
ComPatAI: Artificial-intelligence driven computational pathology, PI Pekka Ruusuvuori, University of Turku, co-PI Teemu Tolonen.
This project will develop a digital pathology workflow to diagnose and grade cancer better. Data availability enables the use of computational approaches for interpreting biopsy images, including modern artificial intelligence-based predictive modeling for diagnostics and grading of cancer. Read more about the topic in this article.
DIGEST: Digital DestinE, PI Reima Eresmaa, FMI, co-PI Heikki Järvinen.
The focus of this project is on the planned Climate and Extreme events Digital Twins in the Destination Earth initiative that will need substantial computing resources. The LUMI installation will offer a unique opportunity for Finland to excel in this initiative. Read more about the topic in this article.
F3AI: Foundation For Finnish Artificial Intelligence, PI Sampo Pyysalo, University of Turku, co-PI Jörg Tiedemann.
This project will create Finnish artificial intelligence models of unprecedented scale and release them freely and openly for any use, thus creating the basis for the next generation of AI systems for Finnish. Read more about the topic & watch a video.
HistoEncoder: A foundation model for all digital histological samples, PI Antti Rannikko, University of Helsinki, co-PI Esa Pitkänen.
This project aims to develop methods for accurate and earlier detection of clinically significant prostate cancer and find correct treatment decisions for an individual patients. The societal objective is to advance health care and increase the patients’ quality of life. A key tool to achieve these aims is digital histopathology, generating massive amounts of imaging data. The research group hopes to improve the accuracy and equality of prostate cancer detection and treatment on a global scale. The research group’s models are trained on pan-cancer data, making them applicable to other cancers as well.
LumiNMT: Multilingual neural machine translation on LUMI scale, PI Jörg Tiedemann, University of Helsinki.
The goal of the project is to train neural machine translation models on a large scale using state-of-the-art transformer models and novel modular multilingual setups. The group focuses on increasing language coverage and efficient use of massively parallel data sets. The research group wants to use LUMI’s extensive parallel computing capabilities to reduce training time and scale up a model size.
TPOST: Transport properties of spherical tokamaks, PI Timo Kiviniemi, Aalto University, co-PI Frank Jenko.
Fusion energy research has matured to the point where it is beginning to attract significant private investment in start-up fusion energy businesses. Recent advances in spherical tokamaks are of particular interest, which optimizes the ratio of volume to the surface to reduce construction costs while simultaneously improving the device’s performance. The simulations performed in this project will help shed light on high-beta plasma turbulence found in the core of the spherical tokamak.