Automated Machine Learning and ML Explainability

This project is building an advanced pipeline for automating machine learning, including model selection, feature selection, hyperparameter tuning and model explainability.

Project Overview

Machine learning (ML) is at the forefront of the rising popularity of data-driven software applications. The resulting rapid proliferation of ML technology, explosive data growth, and shortage of data science expertise have caused the industry to face increasingly challenging demands to keep up with fast-paced develop-and-deploy model lifecycles. Recent academic and industrial research efforts have started to address this problem through automated machine learning (AutoML) pipelines and have focused on model performance as the first-order design objective. Oracle AutoML is a novel iteration-free AutoML pipeline designed to not only provide accurate models, but also in a shorter runtime. These objectives are achieved by eliminating the need to continuously iterate over various pipeline configurations. Our approach to AutoML has been shown to achieve better scores at a fraction of the time compared to state-of-the-art, making it a prime candidate for addressing industry challenges.

Machine learning explainability (MLX) methods provide insight into the behavior of complex machine learning models and the data they work with. MLX helps understand the reasons behind model predictions, can aid debugging, improve performance, identify bias and unfairness in the models, and potentially supports compliance with new regulations. Oracle MLX is focused on building a suite of novel explanation techniques that improve the quality, performance, and interpretability of machine learning explanations for a variety of tasks and applications.

Principal Investigator