Weakly-supervised Contextualization of Knowledge Graph Facts

TL;DR

This paper presents NFCM, a neural method for enriching knowledge graph facts with relevant context by generating and ranking candidate facts, improving user experience in search applications.

Contribution

Introduces NFCM, a supervised learning to rank approach that automatically generates training data for effective KG fact contextualization.

Findings

NFCM significantly outperforms baseline methods in human evaluations.

Automatic training data generation via distant supervision is effective.

The method improves the relevance of contextualized facts in search applications.

Abstract

Knowledge graphs (KGs) model facts about the world, they consist of nodes (entities such as companies and people) that are connected by edges (relations such as founderOf). Facts encoded in KGs are frequently used by search applications to augment result pages. When presenting a KG fact to the user, providing other facts that are pertinent to that main fact can enrich the user experience and support exploratory information needs. KG fact contextualization is the task of augmenting a given KG fact with additional and useful KG facts. The task is challenging because of the large size of KGs, discovering other relevant facts even in a small neighborhood of the given fact results in an enormous amount of candidates. We introduce a neural fact contextualization method (NFCM) to address the KG fact contextualization task. NFCM first generates a set of candidate facts in the neighborhood of a given fact and then ranks the candidate facts using a supervised learning to rank model. The ranking model combines features that we automatically learn from data and that represent the query-candidate facts with a set of hand-crafted features we devised or adjusted for this task. In order to obtain the annotations required to train the learning to rank model at scale, we generate training data automatically using distant supervision on a large entity-tagged text corpus. We show that ranking functions learned on this data are effective at contextualizing KG facts. Evaluation using human assessors shows that it significantly outperforms several competitive baselines.