Question: Matrix multiplication in the DrBC

[gonzalesMK]

Hello!

I've carefully reviewed both your paper and code, and I want to commend you on the excellent work.

I am seeking clarification on a specific aspect, if you would be so kind.

In reference to this specific line of code: https://github.com/FFrankyy/DrBC/blob/b8a84db280e941d27484e338cbfc736774f6a5a6/BetLearn.py#L157C1-L157C58

node_linear = tf.matmul(n2npool, p_node_conv)

It appears to be a matrix multiplication involving weights. However, I'm having difficulty correlating this operation with the visual representation in the provided image from the paper:

Could you please provide me some clarification on this?