Transfer learning: where to transfer from and what to transfer?

• Grant number: 24/09091-3
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: August 01, 2024
• End date: July 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Ana Carolina Lorena
• Grantee: Alfredo Antonio Alencar Exposito de Queiroz
• Host Institution: Divisão de Ciência da Computação (IEC). Instituto Tecnológico de Aeronáutica (ITA). Ministério da Defesa (Brasil). São José dos Campos , SP, Brazil
• Associated research grant: 21/06870-3 - Beyond algorithm selection: meta-learning for data and algorithm analysis and understanding, AP.JP2

Abstract

Transfer learning is a popular strategy to use knowledge extracted from a given source domain to solve another target task. There are many strategies for transfer learning, from sampling part of the data from the source domain to using models induced in a source domain as starting solutions to solve the target tasks. This work will investigate the issue of choosing which source domain can be considered more suitable and what type of knowledge should be transferred for solving a new target task.