Hands-On: AutoGluon

AutoGluon is an open source AutoML framework, developed by AWS, that shattered the SOTA in AutoML on its initial release in 2019 and continues to push the boundaries of what AutoML can achieve. With over two million PyPi downloads, AutoGluon takes inspiration from competition winning solutions on sites like Kaggle and redefines AutoML by ensembling multiple models and stacking them in multiple layers. Among the most challenging of ML problems are datasets that consist of multiple modalities of data, such as text, image, and tabular data. Properly leveraging each modality requires extensive experience and complicated engineering efforts. AutoGluon is able to train on multimodal image-text-tabular data with a single line of code, producing a powerful multi-layer stack ensemble of ResNet image models, BERT language models, and a suite of tabular models all working in tandem. This talk will give an overview of AutoGluon followed by a deep dive into how (and why) it has proven to be so effective, and finish with code examples to demonstrate how you can revolutionize your ML workflow.