Mikhail Voloshin

Learn More
We present a noisy-OR Bayesian network model for simulation-based training, and an efficient search-based algorithm for automatic synthesis of plausible training scenarios from constraint specifications. This randomized algorithm for approximate causal inference is shown to outperform other randomized methods, such as those based on perturbation of the(More)
  • 1