Resources for TensorNetwork

Tensor Network Tutorial by Google X hosted on May 7 2020

Where can I find additional material?

For programming tutorials and additional information visit:

• https://github.com/google/TensorNetwork

Announcements about Tensor Network:

• https://ai.googleblog.com/2019/06/introducing-tensornetwork-open-source.html

• https://blog.tensorflow.org/2020/02/speeding-up-neural-networks-using-tensornetwork-in-keras.html

Summary

Tensor networks are useful for quantum computing, machine learning, and other fields. During the tutorial, you can learn about them and how to use Google X's open-source library called TensorNetwork.

About

X, The Moonshot Factory (formerly Google X) gave a 2 hour virtual session for people looking to get up to speed with tensor networks, a powerful data structure for the optimization of high dimensional datasets in quantum systems and machine learning.

Recently, tensor networks are beginning to find applications in machine learning, with the first clear example in TensorFlow published recently. In the session, X  introduced tensor networks and TensorNetwork --- an open-source library developed by X & Perimeter Institute for efficiently manipulating tensor networks. 

Who is this for?

This tutorial is well suited for students and academics who are involved in computationally intensive tasks across a range of disciplines (computational biology, chemistry, neuroscience, machine learning, quantum physics, etc.). 

What's the agenda?

The first portion was a conceptual introduction to tensor networks and the underlying linear algebra (tensors, vectors, matrices). The second portion was a code walk-through with the TensorNetwork library. Q&A was included.

What do I need to know in advance?

1. Prior Python, tensor network and/or machine learning experience is helpful.  
2. Attendees will need to bring a laptop (nothing to install in advance)
3. These sessions will cover TensorNetwork (NOT TensorFlow). Our goal is to help you get started on TensorNetwork effectively, so you can continue learning and contributing on your own. 

What is the relationship between TensorNetwork and TensorFlow?

TensorNetwork is an open-source library for computation with “Tensor Networks”, a data structure with important applications in quantum computing and machine learning. TensorFlow is an open-source deep learning library. TensorNetwork can be used to implement custom layers for use in TensorFlow programs, and the two libraries work nicely together.