Forgetting to learn logic programs

TL;DR

This paper introduces Forgetgol, a multi-task inductive logic programming system that uses forgetting to improve learning efficiency by reducing hypothesis space and sample complexity, outperforming other methods on large task sets.

Contribution

It presents a novel ILP approach that incorporates forgetting to enhance multi-task learning and demonstrates its effectiveness on extensive task datasets.

Findings

Forgetting reduces hypothesis space and sample complexity.

Forgetgol outperforms other ILP methods on over 10,000 tasks.

Forgetting improves learning efficiency in ILP.

Abstract

Most program induction approaches require predefined, often hand-engineered, background knowledge (BK). To overcome this limitation, we explore methods to automatically acquire BK through multi-task learning. In this approach, a learner adds learned programs to its BK so that they can be reused to help learn other programs. To improve learning performance, we explore the idea of forgetting, where a learner can additionally remove programs from its BK. We consider forgetting in an inductive logic programming (ILP) setting. We show that forgetting can significantly reduce both the size of the hypothesis space and the sample complexity of an ILP learner. We introduce Forgetgol, a multi-task ILP learner which supports forgetting. We experimentally compare Forgetgol against approaches that either remember or forget everything. Our experimental results show that Forgetgol outperforms the alternative approaches when learning from over 10,000 tasks.