I have two matrices that I would like to know how much orthogonal they are with respect to one another. I was wondering if anyone knows a good scoring metric to score the amount of orthogonality between the two matrices?
To explain this more:
I am training a neural net, and I would like to test whether the two layers in the neural net are disentangled. Based on what I read, one way is to measure whether the two layers are learning orthogonal features. And I am looking for a mathematical way of measuring this.