Mapping and Cleaning Open Commonsense Knowledge Bases with Generative Translation

TL;DR

This paper introduces a generative translation method using language models to map and clean open commonsense knowledge bases into fixed schemas, improving accuracy and reducing noise compared to traditional methods.

Contribution

It presents a novel generative translation approach for mapping open KBs into fixed schemas, specifically for commonsense knowledge, balancing accuracy and noise reduction.

Findings

Higher mapping accuracy than rule-based methods

Reduces noise compared to purely generative approaches

Balances traditional and modern KB construction techniques

Abstract

Structured knowledge bases (KBs) are the backbone of many know\-ledge-intensive applications, and their automated construction has received considerable attention. In particular, open information extraction (OpenIE) is often used to induce structure from a text. However, although it allows high recall, the extracted knowledge tends to inherit noise from the sources and the OpenIE algorithm. Besides, OpenIE tuples contain an open-ended, non-canonicalized set of relations, making the extracted knowledge's downstream exploitation harder. In this paper, we study the problem of mapping an open KB into the fixed schema of an existing KB, specifically for the case of commonsense knowledge. We propose approaching the problem by generative translation, i.e., by training a language model to generate fixed-schema assertions from open ones. Experiments show that this approach occupies a sweet spot between traditional manual, rule-based, or classification-based canonicalization and purely generative KB construction like COMET. Moreover, it produces higher mapping accuracy than the former while avoiding the association-based noise of the latter.