Machine Learning for Astrophysics: Proceedings of the ML4Astro International Conference 30 May - 1 Jun 2022 (Astrophysics and Space Science Proceedings, 60)

This book reviews the state of the art in the exploitation of machine learning techniques for the astrophysics community and gives the reader a complete overview of the field. The contributed chapters allow the reader to easily digest the material through balanced theoretical and numerical methods and tools with applications in different fields of theoretical and observational astronomy. The book helps the reader to really understand and quantify both the opportunities and limitations of using machine learning in several fields of astrophysics. Preface Organization Scientific Organizing Committee Local Organizing Committee Referees Sponsoring Institutions and Grants Contents Machine Learning for Hα Emitters Classification 1 Introduction 2 Decision Tree Approach 3 Conclusion References Stellar Dating Using Chemical Clocks and Bayesian Inference 1 Introduction 2 Training Data Sample 3 Inference Technique: Hierarchical Bayesian Model 4 Results 5 Conclusions References Detection of Quasi-Periodic Oscillations in Time Seriesof a Cataclysmic Variable Using Support Vector Machine 1 Introduction 2 Methodology 3 Results, Discussion and Conclusion References Dust Extinction from RF Regression of Interstellar Lines 1 Introduction 2 Methodology 2.1 Automated Equivalent width Measurement 2.2 Random Forest Regression 3 Results References QSOs Selection in Highly Unbalanced Photometric Datasets:The ``Michelangelo'' Reverse-Selection Method 1 Introduction 2 Data Sets 3 QSO Selection References Radio Galaxy Detection Prediction with Ensemble Machine Learning 1 Overview 2 Methodology 3 Results References A Machine Learning Suite to Halo-Galaxy Connection 1 Introduction 2 Methodology 3 Results 4 Discussion and Conclusions References New Applications of Graph Neural Networks in Cosmology 1 Introduction 2 Graph Neural Networks 3 Results 4 Conclusion and Future Prospects References Detection of Point Sources in Maps of the Temperature Anisotropies of the Cosmic Microwave Background References Reconstruction and Particle Identification with CYGNO Experiment 1 The CYGNO Experiment 2 Simulation and Discriminating Variables 3 Models and Training 4 Results and Conclusion References Event Reconstruction for Neutrino Telescopes 1 Neutrino Telescopes 2 Event Reconstruction 2.1 Traditional Approaches 2.2 ML Regression 2.3 Hybrid ML-Likelihood Approaches 3 Conclusions References Classification of Evolved Stars with (Unsupervised) MachineLearning 1 Introduction 2 Methodology 3 Results 4 Conclusions References Patterns in the Chaos: An Unsupervised View of Galactic Supernova Remnants 1 Motivation 2 Methodology 3 Results 4 Conclusions References Clustering of Galaxy Spectra: An Unsupervised Approachwith Fisher-EM 1 Introduction 2 Method 3 Physical Relevance and Robustness of the Classification 4 A Window to Galaxy Evolution References Unsupervised Classification Reveals New Evolutionary Pathways 1 Introduction 2 Data 3 Unsupervised Galaxy Classification 4 Summary References In Search of the Peculiar: An Unsupervised Approachto Anomaly Detection in the Transient Universe 1 Introduction 2 Data 3 The Unsupervised Random Forest 3.1 Feature Selection 4 Discussion References Classifying Gamma-Ray burst X-Ray Afterglows witha Variational Autoencoder 1 Introduction 2 Data and VAE Model 3 Results References Reconstructing Blended Galaxies with Machine Learning 1 Variational Auto-Encoders 2 Data Generation 3 Results References Time Domain Astroinformatics 1 Introduction 2 Synoptic Sky Surveys 3 TDA as a Big Data Challenge 4 Conclusions References A Convolutional Neural Network to Characterise the Internal Structure of Stars 1 The Context 2 The CNN 3 Results and Preliminary Conclusions References Finding Stellar Flares with Recurrent Deep Neural Networks 1 Introduction 2 Flare Detection with Recurrent Deep Neural Networks 3 Searching for Flares with flatwrm2 4 Conclusions References Planetary Markers in Stellar Spectra: Jupiter-Host Star Classification 1 Introduction 2 Data Selection, Preparation and System Design 3 Results 4 Conclusion and Future Work References Using Convolutional Neural Networks to Detect and ConfirmExoplanets 1 Introduction 2 Data and Methods 3 Results 4 Discussion and Conclusions References Machine Learning Applied to X-Ray Spectra: Separating Stars from Active Galactic Nuclei 1 Introduction 2 Data Acquisition and Analysis 3 Results and Discussion 4 Conclusions References Classification of System Variability Using a CNN 1 Introduction 2 Data Simulation and CNN 3 Results and Discussion References Deep Learning Processing and Analysis of Mock Astrophysical Observations 1 Introduction 2 Denoising Radio Interferometric Data 3 X-Ray Source Finder 4 Conclusions References Deep Neural Networks for Source Detection in Radio Astronomical Maps 1 Introduction 2 Dataset: Galactic Radio Images 3 Models 4 Performance Analysis 5 Conclusions References Radio Image Segmentation with Autoencoders 1 Introduction 2 Method 3 Results References Citizen Science and Machine Learning: Towards a Robust Large-Scale Automatic Classification in Astronomy 1 Introduction 2 Methodology 3 Experiments and Results 4 Conclusions References Background Estimation in Fermi Gamma-Ray Burst Monitor Lightcurves Through a Neural Network 1 Introduction 2 Objective 3 Results 4 Conclusion References Machine Learning Investigations for LSST: Strong Lens Mass Modeling and Photometric Redshift Estimation 1 Modeling of Strongly Lensed Galaxies 1.1 Automated Traditional Modeling: glee_ auto.py and glee _ tools.py 1.2 Modeling with Neural Networks 1.3 Direct Comparison Between Neural Network and Traditional Modeling on Real Lenses 2 Photo-z Estimation with a Convolutional Neural Network: NetZ References Multi-Band Photometry and Photometric Redshifts from Astronomical Images 1 Introduction 2 Data 2.1 Training Sample 3 Method 4 Results 5 Conclusions References Inference of Galaxy Clusters' Mass Radial Profiles from Compton-y Maps with Deep Learning Techniques 1 Scientific Context 2 Dataset Selection and Random Forest Training 2.1 Results References Deep Learning 21cm Lightcones in 3D 1 21cm Imaging in 3D to Track the History of the Universe 2 Low-Redshift in 3D: 21cm Source Detection 3 High-Redshift in 3D: 21cm Lightcone Inference 4 Discussion References ConvNets for Enhanced Background Discriminationin the Diffuse Supernova Neutrino Background Search 1 Introduction 2 Diffuse Supernova Neutrino Background 3 Detection and Background 4 Convolutional Neural Network 5 Conclusion and Results References Deep Neural Networks for Single-Line Event Direction Reconstruction in ANTARES 1 Introduction 2 Deep Networks Development and Training 3 Results and Conclusions References Cats vs Dogs, Photons vs Hadrons 1 The Context: Detect Cherenkov Light with ASTRI Telescopes 2 Network and Data 3 Results 4 Summary References Events Classification in MAGIC Through Convolutional Neural Network Trained with Images of Observed Gamma-Ray Events 1 Gamma/Hadron Classification 2 Training the CNN 3 Performance of the CNN and Cross-Checks 4 Conclusion and Future Implication References Federated Learning Meets HPC and Cloud 1 Introduction 2 Federated Learning 3 Hybrid Workflows 4 Hybrid Federated Learning Workflows 5 Conclusion and Future Work References Integration and Deployment of Model Serving Framework at Production Scale 1 Introduction 2 Model Serving 3 Conclusion References Predictive Maintenance for Array of Cherenkov Telescopes 1 Introduction 2 Description of the Scenario 3 Results 4 Conclusions References Index