Well, symbolic AI people also work on probabilistic reasoning. The production level example is ProbLog[1][2], used in genetics. There is even DeepProbLog[3], adding deep learning in the mix. The only problem both implemented in Python. I hope there will be alternatives in native languages. Scryer Prolog[4] might become this implementation one day (it is written in Rust). Another approach is to extend vanilla Prolog, like cplint[5] does.

[1] https://dtai.cs.kuleuven.be/problog/

[2] https://bitbucket.org/problog/problog

[3] https://bitbucket.org/problog/deepproblog

[4] https://github.com/mthom/scryer-prolog

[5] https://github.com/friguzzi/cplint