Machine Learning for Small Bodies in the Solar System

1st Edition - October 29, 2024
Imprint: Elsevier
Editors: Valerio Carruba, Evgeny Smirnov, Dagmara Oszkiewicz
Language: English
Paperback ISBN:
9 7 8 - 0 - 4 4 3 - 2 4 7 7 0 - 5
eBook ISBN:
9 7 8 - 0 - 4 4 3 - 2 4 7 7 1 - 2

Machine Learning for Small Bodies in the Solar System provides the latest developments and methods in applications of Machine Learning (ML) and Artificial Intelligence… Read more

Purchase options

World Book Day Celebration!

Save up to 30% on books and eBooks!

Shop now

Institutional subscription on ScienceDirect

Request a sales quote

Machine Learning for Small Bodies in the Solar System provides the latest developments and methods in applications of Machine Learning (ML) and Artificial Intelligence (AI) to different aspects of Solar System bodies, including dynamics, physical properties, and detection algorithms. Offering a practical approach, the book encompasses a wide range of topics, providing both readers with essential tools and insights for use in researching asteroids, comets, moons, and Trans-Neptunian objects. The inclusion of codes and links to publicly available repositories further facilitates hands-on learning, enabling readers to put their newfound knowledge into practice. Machine Learning for Small Bodies in the Solar System serves as an invaluable reference for researchers working in the broad fields of Solar System bodies; both seasoned researchers seeking to enhance their understanding of ML and AI in the context of Solar System exploration or those just stepping into the field looking for direction on methodologies and techniques to apply ML and AI in their work.

Title of Book
Cover image
Title page
Table of Contents
Copyright
Dedication
Contributors
Foreword
Foreword
Preface
Chapter One: Artificial intelligence and machine learning methods in celestial mechanics
1.1. Introduction
1.2. Machine learning
1.2.1. Supervised learning
1.2.2. Unsupervised learning
1.2.3. Genetic algorithms
1.3. Deep learning
1.4. Conclusions
Chapter Two: Identification of asteroid families' members
2.1. Introduction
2.2. Asteroid families
2.3. Machine learning methods
2.3.1. Machine learning algorithms
2.3.2. Stand-alone methods
2.3.3. Ensemble methods: bagging and boosting classifiers
2.4. Classification of new asteroid families members
2.4.1. Updating family members
2.5. Genetic algorithms
2.5.1. Using genetic algorithms to optimize machine learning predictions
2.6. Final remarks
2.7. Code availability
Chapter Three: Asteroids in mean-motion resonances
3.1. Introduction
3.1.1. Machine-learning methods
3.2. Mean-motion resonances
3.3. Identification of resonant asteroids
3.3.1. Results
3.4. Supervised learning for resonance identification
3.4.1. Challenges of the classical method
3.4.2. Supervised learning
3.4.3. Designations
3.4.4. k-nearest neighbors
3.4.5. Decision tree
3.4.6. Logistic regression
3.4.7. Ensemble learning methods
3.4.8. Naïve Bayes
3.5. Experiments with supervised learning
3.5.1. Comparing the performance of different methods
3.5.2. The role of features
3.5.3. The role of training data volume
3.5.4. Parameters of the best methods
3.5.5. Summary
3.6. Other applications
3.7. Conclusions
3.8. Code availability
Chapter Four: Asteroid families interacting with secular resonances
4.1. Introduction
4.2. Secular resonances: an overview
4.3. Asteroid families interacting with secular resonances
4.3.1. Constraints on the initial ejection velocity field from conserved quantities of secular dynamics
4.3.2. The role of young families, time-reversal numerical methods for dating asteroid families
4.4. Machine learning methods for identifying asteroids interacting with secular resonances: an overview
4.5. ML models based on the proper elements distributions
4.6. Computer vision models
4.6.1. ANN models
4.6.2. CNN models
4.6.3. Model's metrics
4.6.4. The role of regularization
4.6.5. Computer vision applications
4.7. Conclusions and future developments
4.8. Code availability
Chapter Five: Neural networks in celestial dynamics: capabilities, advantages, and challenges in orbital dynamics around asteroids
5.1. Introduction
5.2. Neural networks and their components
5.2.1. FNN and RNN overview
5.2.2. Time series forecasting with neural networks
5.2.3. Applicability to celestial dynamics
5.2.4. Motion relative to (99942) Apophis
5.3. Conclusions
5.4. Code availability
Chapter Six: Asteroid spectro-photometric characterization
6.1. Introduction
6.2. Traditional classification methods
6.3. Overview of ML applications to asteroid spectroscopy and spectro-photometry
6.4. Comparison of ML and traditional methods
6.5. Conclusions and future directions
6.6. Code availability
Chapter Seven: Machine learning-assisted dynamical classification of trans-Neptunian objects
7.1. Introduction to dynamical classification of TNOs
7.1.1. The dynamical classes of TNOs and current approaches to classification
7.1.2. The inherent challenges of identifying resonant TNOs
7.1.3. The need to improve automated classification
7.2. Building a machine learning classifier for TNOs
7.2.1. Building and labeling an adequately large and diverse training dataset
7.2.2. Choosing appropriate and useful time series data features for the classifier
7.2.3. Performance of a gradient boosting classifier for TNO classification
7.3. Looking forward to future applications and improvements
Appendix 7.4. Synthetic TNO orbit generation
Chapter Eight: Identification and localization of cometary activity in Solar System objects with machine learning
8.1. Introduction to the identification of cometary activity in Solar System objects
8.2. Review of classical comet detection methods and strategies
8.2.1. Serendipitous identification of cometary activity of unknown objects
8.2.2. Identification of cometary activity in known objects
8.3. Review of cometary detections methods based on measurements of the individual detections
8.3.1. Identification of comets through spread function parameter analysis
8.3.2. Identification of cometary activity by detection of comae and azimuthal variations
8.3.3. Detection of cometary activity through sporadic, significant deviations from phase curve brightness
8.4. Review of cometary detection methods based on machine learning methods
8.4.1. Identification of cometary objects with convolutional neural networks
8.4.2. Other applications of machine learning used to find comets and future developments
Chapter Nine: Detecting moving objects with machine learning
9.1. Introduction
9.2. Introduction to the detection of moving objects in astronomical imagery
9.2.1. The image search
9.2.2. The linking stage
9.2.3. Digital tracking techniques
9.3. Applications of machine learning
9.3.1. Clustering in moving object detection
9.3.2. Moving object detection and classification in direct images
9.3.3. Trailed moving source detection
9.3.4. Detection of moving objects in shift'n'stack imagery
9.4. Source brightness regression
9.5. Overfitting
9.6. Applications of machine learning in the era of big data surveys
9.7. Code availability
Chapter Ten: Chaotic dynamics
10.1. Introduction
10.2. A brief introduction to chaos in dynamical systems
10.2.1. Chaos in astronomy
10.2.2. Chaotic indicators
10.3. Autocorrelation function indicator (ACFI)
10.3.1. Galaxy dynamics: Henon–Heiles system
10.3.2. Asteroid families
10.3.3. Circular restricted three-body problem
10.4. Conclusions and future developments
10.5. Code availability
Chapter Eleven: Conclusions and future developments
11.1. Introduction
11.2. Assessing the impact of ML applications to small bodies in the Solar System
11.3. Future trends
11.4. Conclusions
Index

Valerio Carruba

Prof. Valerio Carruba is currently an Associate Professor at the São Paulo State University (UNESP) in Brazil. He is one of the founders of the Machine Learning applied to Small Bodies (MASB) research group. His recent interests involve the use of deep learning for the identification of asteroids in secular resonant configurations and machine learning applied for asteroid families’ identification. Asteroid 10741 has been named “Valeriocarruba” by the International Astronomical Union. His paper “Optimization of artificial neural networks models applied to the identification of images of asteroids’ resonant arguments” recently won the CELMEC prize for "Innovative computational methods in Dynamical Astronomy".

Affiliations and expertise

Associate Professor, São Paulo State University (UNESP), Brazil

Evgeny Smirnov

Dr. Evgeny Smirnov works in the field of the dynamics of asteroids. In 2017, he introduced a machine learning approach based on the supervised learning for the identification procedure that decreases the computational time from weeks to seconds. In the same year, he proposed a similar approach for asteroid families instead of the classical HCM method. With a strong background in science and software development, Evgeny connects these areas and brings modern software development patterns and techniques into the field of astronomy.

Affiliations and expertise

Barcelona, Spain

Dagmara Oszkiewicz

Prof. Dagmara Oszkiewicz is a Polish astronomer and planetary scientist. She is an assistant professor at Adam Mickiewicz University in Poznań, Poland, where her research focuses on physical and orbital properties of small Solar System bodies. She has recently expanded her research to include machine learning techniques to the analysis of asteroid spectro-photometric data; her latest work includes applications of machine learning algorithms to the classification of basaltic asteroids in the context of formation of differentiated planetesimals and comparison of various machine learning algorithms for classification of spectro-photometric data from various large sky surveys.

Affiliations and expertise

Assistant Professor, Adam Mickiewicz University in Poznań, Poland

Life Sciences

Physical Sciences & Engineering

Social Sciences & Humanities

Health

Machine Learning for Small Bodies in the Solar System

Purchase options

World Book Day Celebration!

Institutional subscription on ScienceDirect

Valerio Carruba

Evgeny Smirnov

Dagmara Oszkiewicz

Related books