AutoKeras: A Deep Learning AutoML Library
Authors: Haifeng Jin, François Chollet, Qingquan Song, Xia Hu; Published in 2023; Volume 24, Issue 6, Pages 1-6.
Abstract
Deep learning requires expertise in software tools like TensorFlow and Keras, as well as knowledge of model architecture and optimization techniques. This makes it challenging for individuals with limited machine learning and programming experience to leverage deep learning effectively. To address this, we have developed AutoKeras, an Automated Machine Learning (AutoML) library that automates the selection of models and tuning of hyperparameters. AutoKeras provides a simple and accessible interface, allowing users with minimal experience to solve machine learning problems with just a few lines of code. Built on top of Keras and TensorFlow, AutoKeras supports easy export and deployment of models using the TensorFlow ecosystem tools.
[abs]