/awesome-AutoML

Curating a list of AutoML-related research, tools, projects and other resources

GNU General Public License v3.0GPL-3.0

Awesome-AutoML

Curating a list of AutoML-related research, tools, projects and other resources

AutoML

AutoML is the tools and technology to use machine learning methods and processes to automate machine learning systems and make them more accessible. It existed for several decades so it's not a completely new idea.

Recent work by Google Brain and many others have re-kindled the enthusiasm of AutoML and some companies have already commercialized the technology. Thus, it has becomes one of the hosttest areas to look into.

There are many kinds of AutoML, including:

  • Neural network architecture search
  • Hyperparameter optimization
  • Optimizer search
  • Data augmentation search
  • Learning to learn/Meta-learning
  • And many more

Research

AutoML survey

Neural Architecture Search

Neural Architecture Search benchmark

Neural Optimizatizer Search

Activation function Search

AutoAugment

Learning to learn/Meta-learning

Hyperparameter optimization

Automatic feature selection

Model compression

Tools and projects

  • AutoGluon: AutoML Toolkit for Deep Learning
  • hyperunity: A toolset for black-box hyperparameter optimisation
  • auptimizer: An automatic ML model optimization tool
  • Keras Tuner: Hyperparameter tuning for humans
  • Torchmeta: A Meta-Learning library for PyTorch
  • learn2learn: PyTorch Meta-learning Framework for Researchers
  • Auto-PyTorch: Automatic architecture search and hyperparameter optimization for PyTorch
  • ATM: Auto Tune Models: A multi-tenant, multi-data system for automated machine learning (model selection and tuning)
  • Adanet: Fast and flexible AutoML with learning guarantees: Tensorflow package for AdaNet
  • Microsoft Neural Network Intelligence (NNI): An open source AutoML toolkit for neural architecture search and hyper-parameter tuning
  • Dragonfly: An open source python library for scalable Bayesian optimisation
  • H2O AutoML: Automatic Machine Learning by H2O.ai
  • Kubernetes Katib: hyperparameter Tuning on Kubernetes inspired by Google Vizier
  • TransmogrifAI: automated machine learning for structured data by Salesforce
  • Advisor: open-source implementation of Google Vizier for hyper parameters tuning
  • AutoKeras: AutoML library by Texas A&M University using Bayesian optimization
  • AutoSklearn: an automated machine learning toolkit and a drop-in replacement for a scikit-learn estimator
  • Ludwig: a toolbox built on top of TensorFlow that allows to train and test deep learning models without the need to write code
  • AutoWeka: hyperparameter search for Weka
  • automl-gs: Provide an input CSV and a target field to predict, generate a model + code to run it
  • SMAC: Sequential Model-based Algorithm Configuration
  • Hyperopt-sklearn: hyper-parameter optimization for sklearn
  • Spearmint: a software package to perform Bayesian optimization
  • TPOT: one of the very first AutoML methods and open-source software packages
  • MOE: a global, black box optimization engine for real world metric optimization by Yelp
  • Hyperband: open source code for tuning hyperparams with Hyperband
  • Optuna: define-by-run hypterparameter optimization framework
  • RoBO: a Robust Bayesian Optimization framework
  • HpBandSter: a framework for distributed hyperparameter optimization
  • HPOlib2: a library for hyperparameter optimization and black box optimization benchmarks
  • Hyperopt: distributed Asynchronous Hyperparameter Optimization in Python
  • REMBO: Bayesian optimization in high-dimensions via random embedding
  • ExploreKit: a framework forautomated feature generation
  • FeatureTools: An open source python framework for automated feature engineering
  • PocketFlow: use AutoML to do model compression (open sourced by Tencent)
  • DEvol (DeepEvolution): a basic proof of concept for genetic architecture search in Keras

Commercial products

Blog posts

Presentations

Books

Competitions, workshops and conferences

Other curated resources on AutoML

Practical applications