Adaptive Importance Sampling for Bayesian Networks
Note that this currently only works for nodes with two states, although some work has been done to generalize it to networks with more states. This is a (loose) implementation of AIS-BN by Cheng, Druzdzel. See: https://arxiv.org/abs/1106.0253
I couldn't find a code I was happy with, so I wrote this instead. Hope it's useful.