Rule Induction in Knowledge Graphs Using Linear Programming

We present a simple linear programming (LP) based method to learn compact and interpretable sets of rules encoding the facts in knowledge graph (KG) use these solve KG completion problem. Our LP model chooses set bounded complexity from list candidate first-order logic assigns weights them. The bound is enforced via explicit constraints. combine rule generation heuristics with our selection obtain predictions accuracy comparable state-of-the-art codes, even while generating much more sets. Furthermore, when we take as input generated by other often improve interpretability reducing number chosen rules, maintaining accuracy.