Machine learning's widespread adoption is hindered by the challenge of optimizing various aspects, such as hyperparameters, algorithm selection, and preprocessing techniques, for new application tasks. Automated Machine Learning (AutoML) offers a promising solution by assisting developers and researchers in efficiently achieving high-performing models.

In this presentation, we delve into the world of AutoML and discuss research in this field. We explore topics like simultaneous optimization of initial weights, hyperparameters, and algorithms, as well as meta-learning approaches. We will also cover work conducted by our team at CIn.AI, CIn-UFPE, highlighting some contributions made in AutoML.