Development of New Machine Learning Models Based on Gaussian Processes. Applications to Remote Sensing and Astrophysics
- Rafael Molina Soriano Directeur
- Aggelos K. Katsaggelos Directeur/trice
Université de défendre: Universidad de Granada
Fecha de defensa: 05 octobre 2020
- Javier Mateos Delgado President
- Mari Luz García Martínez Secrétaire
- Sandra Morales Martínez Rapporteur
- Juan Gabriel Serra Pérez Rapporteur
- Valeriana Naranjo Ornedo Rapporteur
Type: Thèses
Résumé
In this PhD thesis we have developed different machine learning models based on Gaussian Processes. Different settings (regression, classification and crowdsourcing) are considered, and various application fields (specially remote sensing and astrophysics, but also threat detection and sentiment analysis) are targeted. The main global conclusion of this PhD thesis is the versatility of Gaussian Processes to model different scenarios (regression, classification, crowdsourcing) and target various applications (remote sensing, security, astrophysics), either as the central algorithm to perform the task at hand (Chapters 2-7) or as an auxiliary tool to be integrated within a larger model (Chapter 8)