TensorFlow is a free software library focused on machine learning
created by Google. TensorFlow derives its name from the multidimensional arrays known as tensors, which are used by the neural networks for different operations. TensorFlow is considered the successor of the closed-source application DistBelief and is presently used by Google for research and production purposes. Compared to DistBelief, TensorFlow is faster, smarter and more flexible and can be easily adaptable to new areas and products. It was mainly created for deep neural network research and for facilitating machine learning, though TensorFlow has been used in a wide array of other areas as well. TensorFlow functions by sorting through layers of data (also known as nodes) as part of learning. In the first layer, the system determines the basic features of the object. As deeper movements occur, it looks for more refined information regarding the object. The sorting of images is done at a faster rate, thus giving users more valuable information. TensorFlow is available on different operating systems such as Linux, Windows, MacOS and also on mobile operating platforms like iOS and Android. One of the salient features of TensorFlow is that it is capable of running on multiple CPUs and GPUs. The computations in TensorFlow are reported as stateful dataflow graphs or the Computational Graphs. Currently TensorFlow is used in over six thousand free online repositories.
Computational Graphs:
A computational graph also called the data flow graph, is a directed
graph where the nodes correspond to operations or variables. Variables can feed their value into operations, and operations can feed their output into other operations. This way, every node in the graph defines a function of the variables. The values that are fed into the nodes and come out of the nodes are called tensors, which is just a fancy word for a multi-dimensional array. Hence, it subsumes scalars, vectors and matrices as well as tensors of a higher rank. The concept of a computational graph becomes more useful once the computations become more complex.
Applications Of TensorFlow:
There are many applications using TensorFlow. Many IT giants like
Facebook, Alphabet, Amazon, IBM, Apple, and Microsoft use TensorFlow. Other applications include: a) On-Device Computer Vision for OCR Organization: Google Description: On-device computer vision model to do optical character recognition to enable real-time translation. b) RankBrain Organization: Google Description: A large-scale deployment of deep neural nets for search ranking on www.google.com. c) SmartReply Organization: Google Description: Deep LSTM model to automatically generate email responses. d) Translate Organization: Google Description: It supports over 100 languages translating from one to another. e) Self-driving Cars