Future large scale surveys such as LSST, SKA, JWST and their ongoing pathfinders ZTF, LOFAR, MeerKAT, etc. have ushered astronomy in an era of data-intensive science. The datasets obtained here could span up to exa-scales, and they may contain an unprecedented number of both known and unknown astronomical objects. Machine learning (ML) techniques have been extensively deployed in recent years to mine and classify these objects. Visualisation (VIS) techniques are being developed to provide deeper statistical insights from complex, multi-dimensional datasets. Lessons from the usage of current datasets and ML algorithms are destined to be transferred to future surveys. However, current techniques are more effective in case of uniform noise realisations, low data gaps and well characterised astronomical objects. Further, tremendous data volumes of future surveys demand development of effective real time analyses involving ML and VIS techniques. Such challenges inherent to exa-scale data sizes are unique to astronomy community. Anticipating and resolving them in a timely manner is crucial to maximize scientific yield from future facilities.

In this session we would present cutting-edge techniques at the heart of drawing insight from large-scale data. We aim to: inform newcomers of the current status and ongoing developments in the field, enable researchers to demonstrate relevant software packages, create awareness and recognise future needs of the community in light of current limitations of ML and VIS techniques.

Programme

Invited speakers

• Dr. Anne-Marie Weijmans (Data release coordinator, SDSS-IV/ University of St. Andrews)
• Dr. Ashish Mahabal (Lead Computational and Data Scientist, Caltech)
• Dr. Kalina Borkiewicz (Advanced Visualization Lab, National Center for Supercomputing Applications, University of Illinois at Urbana-Champaign)
...

Scientific organisers

• Dr. Amruta Jaodand, CalTech (US)
• Dr. John Wenskovitch, Virginia Tech (US)
• Martijn Wilhelm (Universiteit Leiden, NL)

Contact

amruta.jaodand @ gmail.com