STEAM: Self-Supervised Taxonomy Expansion with Mini-Paths

TL;DR

STEAM is a self-supervised model for expanding taxonomies by predicting node attachments using mini-paths, significantly improving accuracy over existing methods.

Contribution

The paper introduces STEAM, a novel self-supervised approach leveraging mini-paths and multi-view co-training for taxonomy expansion.

Findings

Outperforms state-of-the-art methods by 11.6% in accuracy.

Achieves 7.0% improvement in mean reciprocal rank.

Effective on three public benchmark datasets.

Abstract

Taxonomies are important knowledge ontologies that underpin numerous applications on a daily basis, but many taxonomies used in practice suffer from the low coverage issue. We study the taxonomy expansion problem, which aims to expand existing taxonomies with new concept terms. We propose a self-supervised taxonomy expansion model named STEAM, which leverages natural supervision in the existing taxonomy for expansion. To generate natural self-supervision signals, STEAM samples mini-paths from the existing taxonomy, and formulates a node attachment prediction task between anchor mini-paths and query terms. To solve the node attachment task, it learns feature representations for query-anchor pairs from multiple views and performs multi-view co-training for prediction. Extensive experiments show that STEAM outperforms state-of-the-art methods for taxonomy expansion by 11.6\% in accuracy and 7.0\% in mean reciprocal rank on three public benchmarks. The implementation of STEAM can be found at \url{https://github.com/yueyu1030/STEAM}.