6 ms·
I am curious to know if this type of network can help with causal inference.
by nu91 2y ago
I am curious to know if this type of network can help with causal inference.
- esafak 2y agoThey help with interpretability, which is a step brother. See for example "From Shapley Values to Generalized Additive Models and back" https://arxiv.org/abs/2209.04012 https://arxiv.org/abs/2209.04012
- nu91 2y agoIndeed, Thank you!
- fermisea 2y agoThey definitely do, I'm planning to release a new package to do that in a couple weeks