SKA Science Data Challenge 2: analysis and results

Abstract

The Square Kilometre Array Observatory (SKAO) will explore the radio sky to new depths in order to conduct transformational science. SKAO data products made available to astronomers will be correspondingly large and complex, requiring the application of advanced analysis techniques to extract key science findings. To this end, SKAO is conducting a series of Science Data Challenges, each designed to familiarise the scientific community with SKAO data and to drive the development of new analysis techniques. We present the results from Science Data Challenge 2 (SDC2), which invited participants to find and characterise 233245 neutral hydrogen (Hi) sources in a simulated data product representing a 2000~h SKA MID spectral line observation from redshifts 0.25 to 0.5. Through the generous support of eight international supercomputing facilities, participants were able to undertake the Challenge using dedicated computational resources. Alongside the main challenge, `reproducibility awards' were made in recognition of those pipelines which demonstrated Open Science best practice. The Challenge saw over 100 participants develop a range of new and existing techniques, with results that highlight the strengths of multidisciplinary and collaborative effort. The winning strategy -- which combined predictions from two independent machine learning techniques to yield a 20 percent improvement in overall performance -- underscores one of the main Challenge outcomes: that of method complementarity. It is likely that the combination of methods in a so-called ensemble approach will be key to exploiting very large astronomical datasets.

arXiv commentUnder review by MNRAS; 28 pages, 16 figures

References in corpus (49)

• Planck 2015 results. XIII. Cosmological parameters
• scikit-image: Image processing in Python
• Attention U-Net: Learning Where to Look for the Pancreas
• TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation
• U${}^2$-Net: Going Deeper with Nested U-Structure for Salient Object Detection
• ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed data
• THINGS: The HI Nearby Galaxy Survey
• The Baryonic Tully-Fisher Relation
• Cold gas accretion in galaxies
• On the Compactness, Efficiency, and Representation of 3D Convolutional Networks: Brain Parcellation as a Pretext Task
• The cold gaseous halo of NGC 891
• New Lessons from the HI Size-Mass Relation of Galaxies
• SoFiA: a flexible source finder for 3D spectral line data
• Deep H I Survey of the Spiral Galaxy NGC 2403
• WALLABY -- An SKA Pathfinder HI Survey
• The Westerbork Hydrogen Accretion in Local Galaxies (HALOGAS) Survey I. Survey Description and Pilot Observations
• The ALFALFA HI mass function: A dichotomy in the low-mass slope and a locally suppressed 'knee' mass
• Predictions for ASKAP Neutral Hydrogen Surveys
• The Strong Gravitational Lens Finding Challenge
• SoFiA 2 -- An automated, parallel HI source finding pipeline for the WALLABY survey
• Cosmological evolution of atomic gas and implications for 21 cm HI absorption
• The Tiered Radio Extragalactic Continuum Simulation (T-RECS)
• The Arecibo Dual-Beam Survey: The HI Mass Function of Galaxies
• Highest Redshift Image of Neutral Hydrogen in Emission: A CHILES Detection of a Starbursting Galaxy at z=0.376
• Reinventing 2D Convolutions for 3D Images
• Galaxy formation in the Planck Millennium: the atomic hydrogen content of dark matter halos
• A maximum-likelihood method for improving faint source flux and color estimates
• Tracing HI Beyond the Local Universe
• Anticipated Performance of the Square Kilometre Array -- Phase 1 (SKA1)
• The Arecibo Dual-Beam Survey: Arecibo and VLA Observations
• The Exploration of the Unknown
• The Redshift Evolution of the Mass Function of Cold Gas in Hierarchical Galaxy Formation Models
• Atomic hydrogen in star-forming galaxies at intermediate redshifts
• Square Kilometre Array Science Data Challenge 1: analysis and results
• DeepSource: Point Source Detection using Deep Learning
• A Hybrid Ensemble Learning Approach to Star-Galaxy Classification
• Giant Metrewave Radio Telescope Detection of HI 21 cm Emission from Star-forming Galaxies at $z\approx 1.3$
• Transfer learning for radio galaxy classification
• An investigation on the factors affecting machine learning classifications in $\gamma$-ray astronomy
• Stacking for machine learning redshifts applied to SDSS galaxies
• Stacking machine learning classifiers to identify Higgs bosons at the LHC
• A Unified Framework for Constructing, Tuning and Assessing Photometric Redshift Density Estimates in a Selection Bias Setting
• The Characterised Noise Hi source finder: Detecting Hi galaxies using a novel implementation of matched filtering
• 2D-1D Wavelet Reconstruction As A Tool For Source Finding In Spectroscopic Imaging Surveys
• The SKA view of the Neutral Interstellar Medium in Galaxies
• Lightweight HI source finding for next generation radio surveys
• A comparative study of source-finding techniques in HI emission line cubes using SoFiA, MTObjects, and supervised deep learning
• A Framework for HI Spectral Source Finding Using Distributed-Memory Supercomputing
• A Fresh Look at FAIR for Research Software

Cited by in corpus (12)

• A review of unsupervised learning in astronomy
• The Tiered Radio Extragalactic Continuum (T-RECS) simulation II: HI emission and continuum-HI cross-correlation
• YOLO-CIANNA: Galaxy detection with deep learning in radio data. I. A new YOLO-inspired source detection method applied to the SKAO SDC1
• Optimal, fast, and robust inference of reionization-era cosmology with the 21cmPIE-INN
• SKATR: A Self-Supervised Summary Transformer for SKA
• The Power of the Cosmic Web
• Deep Learning 21cm Lightcones in 3D
• The Galaxy Activity, Torus, and Outflow Survey (GATOS). Black hole mass estimation using machine learning
• On the completeness and reliability of visual source extraction : an examination of eight thousand data cubes by eye
• ComPACT: ACT+Planck galaxy cluster catalogue
• Bringing computation to the data: A MOEA-driven approach for optimising data processing in the context of the SKA and SRCNet
• WALLABY Pilot Survey: HI source-finding with a machine learning framework