Data science is emerging as a new development in astronomy and computational science. Modern astronomical observations produce extremely large data volumes. As astrophysics and cosmology are rich with data, even comparable to the data volumes in high energy physics, nowadays we are looking into development and application of new data-mining technologies. In this way we can fully embracing modern astronomy as a data rich science. In particular Machine Learning became a popular method to analyze astronomical data.
Kapteyn staff is involved with data science through several projects, collaborating closely with institutes all over Europe. Since 2017, the main projects incorparating Data Science is SUNDIAL (SUrvey Network for Deep Imaging Analysis & Learning). Prof. dr. Reynier Peletier is the SUNDIAL, network coordinator. In addition 6 PhD students of the Kapteyn astronomical institute are part of the Data Science & Systems Complexity Centre.
Researchers in Computational Intelligence (CI) and Astronomy are merging their knowledge by forming a consortium, SUNDIAL. SUNDIAL is a unique intensive collaboration as it forms an interdisciplinary network of nine research groups in The Netherlands, Germany, Finland, France, the United Kingdom, Spain, Belgium and Italy. It creates a platform for broad and data science intensive research.
It aims to analyse the ever growing large surveys in astronomy, done by studing galaxy formation and their evolution over the history of time. Astronomers will learn how to uncover structures hidden in the data. Without the help of CI, these structures would have remain masked. This is done by combining two radically different ways to approch data: purely data-driven machine learning and specialist approaches based on techniques developed in astronomy.
In short, the aim of SUNDIAL is to train researchers to address the most prominent CI topics related to the analysis of Big Data and their application to galaxy evolution studies. To acchieve this objective, the focus lies on the following topics:
- Automatisation of faint low surface brightness features (dwarf galaxies, merger remnants, intracluster light) in deep astronomical surveys.
- Astrophysically interpreting these features in terms of galaxy formation and evolution.
- Automatisation of object recognition in astronomical data sets using:
- Clustering: an unsupervised identification technique to identify groups of similar objects (showing similar properties)
- Classification: an supervised technique to assign objects to predefined classes.
- Prior information obtained from astrophyics is crusial to deliver the required results.
- Simulating the characterisation, visualisation and the interaction between galaxies. Using the simulations we want to idenify cristical characterisations. Comparing the results will lead to a better understanding of the evolution of a.o. galaxy clusters.
Training young researchers though a combination of astrophysics and computer science:
Data Science & Systems Complexity (DSSC) centre
The Kapteyn Astronomical Institute is stronly involved in the DSSC, as 6 of the 14 PhD students are affiliated with the institute. The DSSC Centre aims to understand and design complex systems and processes through massive data by bringing together over 70 researchers from several disciplines with an immediate interest in the handling of Big Data and Complexity. The goals of the collaboration is to
- conduct fundamental and applied scientific research in the field of data science and complex systems, which is both long-lasting and cross-disciplinary,
- serve as a platform where students can apply to for their education in data and complexity science,
- and educate a new generation of scientists and professionals to be successful in their own career.
More information can be found on the website of the DCCS.
|Last modified:||18 October 2019 3.44 p.m.|