Towards Probabilistic Inductive Logic Programming with Neurosymbolic Inference and Relaxation

TL;DR

This paper introduces Propper, a novel probabilistic ILP framework that integrates neurosymbolic inference and relaxation techniques, enabling learning from noisy, probabilistic data with minimal examples, outperforming existing models.

Contribution

Propper extends ILP to handle probabilistic background knowledge using neurosymbolic inference, BCE, and a relaxed hypothesis constrainer, advancing learning from noisy sensory data.

Findings

Propper learns from as few as 8 examples.

Propper outperforms binary ILP and Graph Neural Networks.

Propper effectively handles flawed and probabilistic background knowledge.

Abstract

Many inductive logic programming (ILP) methods are incapable of learning programs from probabilistic background knowledge, e.g. coming from sensory data or neural networks with probabilities. We propose Propper, which handles flawed and probabilistic background knowledge by extending ILP with a combination of neurosymbolic inference, a continuous criterion for hypothesis selection (BCE) and a relaxation of the hypothesis constrainer (NoisyCombo). For relational patterns in noisy images, Propper can learn programs from as few as 8 examples. It outperforms binary ILP and statistical models such as a Graph Neural Network.