Generative adversarial networks on chaotic systems

• Grant number: 23/04172-2
• Support Opportunities: Scholarships abroad - Research Internship - Master's degree
• Start date: July 01, 2023
• End date: December 31, 2023
• Field of knowledge: Physical Sciences and Mathematics - Physics
• Principal Investigator: Odemir Martinez Bruno
• Grantee: João Pedro do Valle Alvarenga
• Supervisor: Bernard de Baets
• Host Institution: Instituto de Física de São Carlos (IFSC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Institution abroad: Ghent University (UGent), Belgium
• Associated to the scholarship: 22/01935-2 - Patterns and pseudorandomness in iterative maps in a chaotic regime, BP.MS

Abstract

The objective of this project is to train the candidate in the development of Generative Adversarial Networks (GANs) to generate time series with chaotic characteristics. The project will be developed in the research group of Prof. Dr. Bernard De Beats, installed at Ghent University. In this research group, the candidate will implement state-of-the-art methods of GANs for time series, exploring the most diverse types of neural network architectures and different chaotic orbits. To validate the GANs, classic metrics of time series and visualization methods of dynamical systems will be applied. In this way, we can achieve time series with local and global characteristics of chaotic behavior. The knowledge acquired by the candidate during the project in Belgium will serve to advance the research project related to the fellowship in Brazil. (AU)