Machine Intelligence at Google Scale
1) Google uses neural networks and deep learning across many of its services like Search, Photos, Translate, and Android apps. 2) Google provides external access to machine learning through APIs like Cloud Vision, Speech, Translation and Natural Language that allow developers to easily integrate ML into applications. 3) TensorFlow is Google's open source library for machine learning that makes it easy to design, train and deploy models at scale. 4) Google trains models using distributed processing on thousands of GPUs in its datacenters and also provides Cloud ML to allow external users to train models in the cloud.