TensorFlow open-source software library for Machine Learning
We are blown away by what we can achieve in the sphere of image recognition!
TensorFlow™ is an open source software library for high-end computation developed following many years of work by researchers and engineers working on the Google Brain Team within Google's Machine Intelligence research organisation and is the Artificial Intelligence technology which powers Google Photos to Google Cloud Speech.
Released on 14 November 2017 as a "developer release" (i.e. bleeding-edge but work in progress), the TensorFlow Lite API allows Android apps on smartphones to accomplish Machine Learning.
The flexible architecture allows us to deploy computation to one or more CPUs or GPUs in a desktop, server, or mobile device with a single API.
Machine Learning will change the computing landscape radically over the coming months and years, especially as users become accustomed to voice operated intelligent personal assistants such as Amazon Alexa/Echo and Google Home where thay have human-like interactions with their devices via the camera and voice interaction.
Whilst it is easy for a human to recognise a human’s face and to tell the difference between a photo of a panther or a cat, these are actually very difficult tasks for a computer to perform accurately. The science in involved in Machine Learning has raced ahead in the last few years and a kind of Machine Learning model called a “deep convolutional neural network” can be trained to perform complex visual recognition tasks which and which can out-perform humans with regard to both accuracy and speed.
We have been testing the use of artificial intelligence and machine learning for our bespoke Android apps and we are simply blown away by what we can achieve in the sphere of image recognition by training deep convolutional neural networks with machine learning algorithms.