DeepProbLog: Neural Probabilistic Logic Programming
I will talk about DeepProbLog, which was previously published at NeurIPS 2018. DeepProbLog is an extension of Problog, a probabilistic logic programming language (PLP) that incorporates neural networks through the neural predicates. This neural predicate represents the relation between the input and output as defined by a neural network. I will demonstrate the capabilities of DeepProbLog in combined symbolic and subsymbolic reasoning, program induction, and probabilistic logic programming.
I’m a PhD student in the DTAI research group at KU Leuven. My research is centered around neuro-symbolic integration, more specifaclly, in intergrating neural networks into probabilistic logic programming.