Amateur here: How can we write a 2D transposed convolution (aka deconvolution) using the steepest descent method given the following restrictions:
cannot use any Python built-in functions
cannot explicitly transpose the matrix
cannot explicitly perform matrix-vector multiplications
I have already written a forward 2D convolution method, and would like to reuse it to perform the backward process. I am not sure how this can be done with the given restrictions.
Zero-padding the filters did not seem to work...
I don't have a specific background in data science so any help is appreciated, thanks!