Google has released AdaNet to make machine learning even easier.
Google today introduced AdaNet, an open source tool that combines machine-learning algorithms for better estimates.
Google’s artificial intelligence engineer Charles Weil said, “AdaNet is here to make our AutoML work faster. More importantly, decentralization not only provides a general framework that combines learning to learn neural network structure but also better models.”
Adanet-uses an approach called collective learning to combine and develop algorithms. This method required domain expertise or a lot of time before.
To make this process even easier, the framework connects to the TensorFlow Estimator to collect important data somewhere. In addition, while the artificial intelligence model is being trained, the TensorBoard sends visual feedback.
By learning about the structures of neural networks, AdaNet offers a learning guarantee for collective models. Then it adds subnets to these networks.
Anyone who implements machine learning and wants to have more control over the process can use the TensorFlow application programming interfaces (API) to identify their own subnets, edit lost functions, or adjust other settings.