A Generative AutoML for Tabular Data

An Automated-Machine Learning (AutoML) platform aims to automate the process of data engineering, feature engineering, hyper-parameter optimization, training, prediction, and deployment of a model, where it minimizes human supervision in all stages. One of the popular aspects of artificial intelligence utilization across different domains is AutoML for tabular data (structured data). In this talk, I will focus on the technology behind AutoML tools for tabular data and I will discuss advantages and limitations of this technology. Finally, I will conclude this session by demonstrating Fujitsu’s AutoML platform that produces/deploys a machine learning pipeline for tabular datasets across different domains. I will highlight how the tool generates a pipeline source-code with explanations that introduce transparency, flexibility and explainability of the recommended machine-learning models.

View the slides for this session