Think of TensorFlow as a high-performance, easy-to-use framework to perform machine learning. The specific algorithm used for this machine learning is neural networks. But these are not ordinary neural nets. These are a souped-up version, including fancy stuff such as convolutional or adversarial neural nets. Created by the Google Brain team, TensorFlow provides a user-friendly python front-end, but the computations are done in high performance C++ binaries. Can you tell how excited I am to learn more about this framework and share it with you (if you can’t, I need to work on my writing skills!)?
There are better written resources on the web. Here are a few I liked:
- From the developers themselves: TensorFlow
- InfoWorld: What is TensorFlow? The machine learning library explained
- Play around with a NeuralNet, right in your browser!