Hands-On Reinforcement Learning with Python

Computation graph

Everything in TensorFlow will be represented as a computational graph that consists of nodes and edges, where nodes are the mathematical operations, say addition, multiplication and so on, and edges are the tensors. Having a computational graph is very efficient in optimizing resources and it also promotes distributed computing.

Say we have node B, whose input is dependent on the output of node A; this type of dependency is called direct dependency.

For example:

A = tf.multiply(8,5)B = tf.multiply(A,1)

When node B doesn't depend on node A for its input it is called indirect dependency.

For example:

A = tf.multiply(8,5)B = tf.multiply(4,3)

So if we can understand these dependencies, we can distribute the independent computations ...