Hello Pymc community >.<
I am fairly new to the world of Bayesian modeling and recetly I am trying to find a proper way of constructing a Bayesian optimization model using non-guassian process (GP) surrogate models. I found Baysian neural nets could potentially be a good candidate to replace GP as the surrogate model in Bayesian optimization. I am wondering if anyone has the experimence of trainig a Bayesian optimization with Bayesian neural nets? Dose Pymc support this, or other libs should be involved? Any suggestions are appreciated!
Thanks in advance!