2 Hits in 2.4 sec

Structural Causal Models Are (Solvable by) Credal Networks [article]

Marco Zaffalon and Alessandro Antonucci and Rafael Cabañas
2020 arXiv   pre-print
This contribution should be regarded as a systematic approach to represent structural causal models by credal networks and hence to systematically compute causal inferences.  ...  This allows to exactly map a causal model into a credal network.  ...  The tests have been performed by means of a Java library implementing all the techniques discussed in the paper.  ... 
arXiv:2008.00463v1 fatcat:kenlbdt5qjfwljgnjacwcr6iha

Robust Generalizations of Stochastic Derivative-Free Optimization

Julian Rodemann
Two of them, the generalized lower confidence bound and a weighted maximum likelihood approach, outperform standard Bayesian optimization for specific types of problems.  ...  Feasibility Study Implementations of "ApproxLP" and "CVE" are part of the publicly available java library CREDICI (CREDal Inference for Causal Inference) [Zaffalon et al., 2020] .  ...  In a recent work, [Zaffalon et al., 2020] show that structural causal models can be solved by established algorithms for inference in credal nets [Augustin et al., 2014, Chapter 9 .5] such as "ApproxLP  ... 
doi:10.5282/ubm/epub.77441 fatcat:strgnvbi5fbgvkwq4nuw756tfe