Intelligent Astrophysics (Emergence, Complexity and Computation, 39)

✍ Scribed by Ivan Zelinka (editor), Massimo Brescia (editor), Dalya Baron (editor)

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 300
• Edition: 1st ed. 2021
• Category: Library

✦ Synopsis

This present book discusses the application of the methods to astrophysical data from different perspectives. In this book, the reader will encounter interesting chapters that discuss data processing and pulsars, the complexity and information content of our universe, the use of tessellation in astronomy, characterization and classification of astronomical phenomena, identification of extragalactic objects, classification of pulsars and many other interesting chapters. The authors of these chapters are experts in their field and have been carefully selected to create this book so that the authors present to the community a representative publication that shows a unique fusion of artificial intelligence and astrophysics. 

✦ Table of Contents

Preface
Contents
Artificial Intelligence in Astrophysics
1 Introduction
2 Artificial Intelligence and Machine Learning
2.1 Brief Overview of Artificial Intelligence
2.2 Learning Algorithms
2.3 Machine Learning: An Overview
2.4 Machine Learning Process
2.5 Machine Learning Algorithms
2.6 Evaluation Metrics
3 EA and Swarm Overview
3.1 Evolutionary Algorithm
3.2 Swarm Intelligence
3.3 Examples
4 Limits to Computation
4.1 Searched Space and Its Complexity
4.2 Physical Limits of Computation
5 Stellar Data Classification
6 Data Preprocessing
7 Classification Method
8 Conclusion
References
The Complexity and Information Content of Simulated Universes
1 Introduction
2 Information and Complexity: An Overview
2.1 Shannon's Information Entropy
2.2 The Algorithmic Complexity
2.3 The Statistical Complexity
2.4 The Block Entropy and the Entropy Gain
2.5 The Excess Entropy and the Efficiency of Prediction
3 Results
3.1 How Complex Is the Formation of a Galaxy Cluster?
3.2 How Complex Is the Formation of the Cosmic Web?
4 Conclusions
References
The Voronoi Tessellation Method in Astronomy
1 The Voronoi Tessellation in a Spatial Galaxy Distribution: First Works and Basic Approach
2 Voronoi Tessellation of the First, Second and Third Orders: Identification of the Low-Populated Galaxy Groups, Environment Effect, and Dark Matter Content
3 The Voronoi Tessellation in Astrophysical Research
4 The Voronoi Tessellation and Machine Learning
5 Instead of Conclusion
References
Statistical Characterization and Classification of Astronomical Transients with Machine Learning in the era of the Vera C. Rubin Observatory
1 Introduction
2 Data
2.1 The SNPhotCC Simulated Catalogue
2.2 The PLAsTiCC Simulated Catalogue
2.3 The Statistical Parameter Space
3 Machine Learning Models
3.1 The Random Forest Classifier
3.2 The Nadam, RMSProp and Adadelta Classifiers
3.3 Parameter Space Exploration
3.4 Classification Statistics
4 Experiments
4.1 Data Pre-processing
4.2 Periodic Versus Non Periodic
4.3 Handling of Negative Fluxes
4.4 Optimization of the Parameter Space for Transients
4.5 Supernovae Versus All
4.6 Supernovae Ia Versus II
4.7 Superluminous SNe Versus SNe I
4.8 Simultaneous Classification of Six SNe Sub-Types
5 Discussion and Conclusions
References
Application of Machine and Deep Learning Methods to the Analysis of IACTs Data
1 Introduction
2 The Simulations
3 The Muon Case
3.1 Image Cleaning Method
3.2 Choice of the Parameters
3.3 Architecture and Results
3.4 Transfer Learning for the Muon Case
3.5 Experimental Results of Transfer Learning over the Muon Case
4 The Gamma / Hadron Case
4.1 Models
5 Results
6 Conclusions
References
Intelligent Photometric Identification of Extragalactic Objects from AllWISEtimesPan-STARRS DR1 Data
1 Introduction
2 Data and Sample Selection
2.1 AllWISE
2.2 Pan-STARRS DR1
2.3 Positional Cross-Matching
2.4 Training Sample: SDSS DR14
3 Classification Model
3.1 Formal Description
3.2 Technical Details
4 Application of Approach
4.1 Classification Result
5 Validation with External Data
5.1 Spectroscopic Datasets
5.2 Validation with Galaxy Catalogues
5.3 Astrometric Validation
6 Conclusions
References
Ensemble Classifiers for Pulsar Detection
1 Introduction
2 Background
3 Imbalanced Classification
3.1 A Closer Look at Rarity
3.2 Addressing Imbalance—Sampling
3.3 Addressing Imbalance—Cost-Sensitive Classification
4 Ensemble Classification
4.1 Rationale Behind Ensembles
4.2 Voting
4.3 Bagging
4.4 Boosting
5 Ensembles for Imbalanced Classification
5.1 SMOTEBagging
5.2 SMOTEBoost
5.3 EasyEnsemble
5.4 Balanced Random Forest
5.5 AdaC2
5.6 AdaCost
5.7 USBE
6 Experimental Results
6.1 Dataset
6.2 Performance Measures
6.3 Results
7 Conclusions
References
Periodic Astrometric Signal Recovery Through Convolutional Autoencoders
1 Introduction
2 Astrometry
2.1 Narrow-Angle Astrometry
3 TOLIMAN
3.1 The TOLIMAN Data Challenge
4 Simulations
4.1 The Fast Fourier Transform
4.2 Generating Data
5 Dimensionality Reduction
6 Deep Learning
6.1 Convolutional Layer
6.2 Pooling Layer
6.3 Deep Convolutional Autoencoder
7 Model Architecture
8 Signal Analysis
8.1 The Lomb-Scargle Periodogram
8.2 Atom Time Series Analysis
9 Experiments and Results
9.1 Discussion of Results
10 Conclusions
References
Comparison of Outlier Detection Methods on Astronomical Image Data
1 Introduction
2 The Data
2.1 Data Exploration and Preprocessing
3 Outlier Detection Methods
3.1 PCA-Based Detection Methods
3.2 Autoencoders
4 Discussion
4.1 Future Work
References
Anomaly Detection in Astrophysics: A Comparison Between Unsupervised Deep and Machine Learning on KiDS Data
1 Introduction
2 Data Preparation
3 Disentangled Convolutional Autoencoders
3.1 Validation with Synthetic Data
3.2 Application to KiDS Data
4 Unsupervised Random Forests
4.1 Anomaly Detection in KiDS Data Based on the URF
5 Discussion
6 Conclusions
References
Rejection Criteria Based on Outliers in the KiDS Photometric Redshifts and PDF Distributions Derived by Machine Learning
1 Introduction
2 Probability Density Function
3 Data
4 Methods
5 Results
5.1 Zphot and Stacked PDF Statistics
5.2 PIT and Credibility Analysis
6 Conclusions
References
Large Astronomical Time Series Pre-processing for Classification Using Artificial Neural Networks
1 Introduction
1.1 Types of Stars Variability
2 State of the Art
2.1 Time Series Classification Methods
2.2 Other Related Work
3 Data Sets
3.1 BRITE
3.2 Kepler K2
4 Artificial Neural Networks Approaches
4.1 Data Pre-processing
5 Experiments and Results
5.1 BRITE Data Set
5.2 Kepler K2 Data Set
6 Conclusions
References
Frontiers