Optimal transport for machine learning: theory and applications

• Ano de defesa: 2021
• Autor(a) principal: Barreira, Davi Sales
• Orientador(a): Mendes, Eduardo Fonseca
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Palavras-chave em Português: Distância de Wasserstein
• Palavras-chave em Inglês: Optimal transport; Wasserstein distance; Machine learning; Literature review
• Link de acesso: https://hdl.handle.net/10438/30407

O que os operadores de produção de petróleo valorizam ao comprar produtos químicos?: uma análise sobre a percepção de valor na decisão de compra ou contratação de um provedor de especialidades químicas no mercado de óleo e gásIn recent years, advances in Optimal Transport have led to a surge of applications in fields such as Economics, Quantitative Finance and Signal Processing, among others. One area in which it has been found particularly successful is Machine Learning. The development of computationally efficient methods for solving Optimal Transport problems opened doors for creating Machine Learning algorithms using concepts from Optimal Transport. These new algorithms encompass many different sub-areas such as Transfer Learning, Clustering, Dimensionality Reduction, Generative Models, just to name some. This work provides an overview of the different ways in which Optimal Transport has been used in Machine Learning, thus helping Machine Learning researchers to better understand its impact in the field and how to use it. This thesis first introduces the main theoretical and computational aspects of Optimal Transport theory in an accessible way to Machine Learning researchers, followed by a semi-systematic literature review focusing on the main uses of Optimal Transport in Machine Learning.