AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data

Abstract

We introduce AutoGluon-Tabular, an open-source AutoML framework that requires only a single line of Python to train highly accurate machine learning models on an unprocessed tabular dataset such as a CSV file. Unlike existing AutoML frameworks that primarily focus on model/hyperparameter selection, AutoGluon-Tabular succeeds by ensembling multiple models and stacking them in multiple layers. Experiments reveal that our multi-layer combination of many models offers better use of allocated training time than seeking out the best. A second contribution is an extensive evaluation of public and commercial AutoML platforms including TPOT, H2O, AutoWEKA, auto-sklearn, AutoGluon, and Google AutoML Tables. Tests on a suite of 50 classification and regression tasks from Kaggle and the OpenML AutoML Benchmark reveal that AutoGluon is faster, more robust, and much more accurate. We find that AutoGluon often even outperforms the best-in-hindsight combination of all of its competitors. In two popular Kaggle competitions, AutoGluon beat 99% of the participating data scientists after merely 4h of training on the raw data.

References in corpus (9)

• Scikit-learn: Machine Learning in Python
• AutoML: A Survey of the State-of-the-Art
• CatBoost: gradient boosting with categorical features support
• fastai: A Layered API for Deep Learning
• Self-Normalizing Neural Networks
• Entity Embeddings of Categorical Variables
• Towards Automated Machine Learning: Evaluation and Comparison of AutoML Approaches and Tools
• OBOE: Collaborative Filtering for AutoML Model Selection
• An Open Source AutoML Benchmark

Cited by in corpus (30)

• Best Practices for Scientific Research on Neural Architecture Search
• Automated Evolutionary Approach for the Design of Composite Machine Learning Pipelines
• Automated Machine Learning on Graphs: A Survey
• FLAML: A Fast and Lightweight AutoML Library
• PHOTONAI -- A Python API for Rapid Machine Learning Model Development
• Is deep learning necessary for simple classification tasks?
• Weight-Sharing Neural Architecture Search: A Battle to Shrink the Optimization Gap
• Fast, Accurate, and Simple Models for Tabular Data via Augmented Distillation
• DARWIN: A Highly Flexible Platform for Imaging Research in Radiology
• Enabling Collaborative Data Science Development with the Ballet Framework
• Auto-PyTorch Tabular: Multi-Fidelity MetaLearning for Efficient and Robust AutoDL
• Multi-objective Asynchronous Successive Halving
• Amazon SageMaker Autopilot: a white box AutoML solution at scale
• VEGA: Towards an End-to-End Configurable AutoML Pipeline
• LightAutoML: AutoML Solution for a Large Financial Services Ecosystem
• AgEBO-Tabular: Joint Neural Architecture and Hyperparameter Search with Autotuned Data-Parallel Training for Tabular Data
• Understanding Prediction Discrepancies in Machine Learning Classifiers
• How to set up your first machine learning project in astronomy
• A Neophyte With AutoML: Evaluating the Promises of Automatic Machine Learning Tools
• Ensemble Squared: A Meta AutoML System
• Simple Modifications to Improve Tabular Neural Networks
• Benchmarking Multimodal AutoML for Tabular Data with Text Fields
• Cocktail: Leveraging Ensemble Learning for Optimized Model Serving in Public Cloud
• Domain Knowledge Injection in Bayesian Search for New Materials
• Automatic Componentwise Boosting: An Interpretable AutoML System
• Distiller: A Systematic Study of Model Distillation Methods in Natural Language Processing
• AutoML Meets Time Series Regression Design and Analysis of the AutoSeries Challenge
• Surrogate Model Based Hyperparameter Tuning for Deep Learning with SPOT
• AutoSmart: An Efficient and Automatic Machine Learning framework for Temporal Relational Data
• Man versus Machine: AutoML and Human Experts' Role in Phishing Detection