I have a question about the various filters that can be applied to a GLM. I understand the high level idea of what we’re trying to do, but how do you get those filters? Are they something that we create based on a priori hypotheses about what’s going on (and if so, how do we get the math to do that), or do we get those via the optimization processes we were practicing in the various tutorials? If we get them through the optimization process, how do we relate the parameters from optimization to those temporal filters?
The filters are the thetas found. You can see this in the tutorial where, after optimizing, we plot the thetas and it looks like the filters discussed on the slides.
1 Like