Alleviating Sparsity of Open Knowledge Graphs with Ternary Contrastive Learning

TL;DR

This paper introduces TernaryCL, a contrastive learning framework that leverages ternary propagation patterns to address sparsity and zero-shot challenges in Open Knowledge Graphs, significantly improving entity and relation representations.

Contribution

TernaryCL is the first contrastive learning approach that models ternary propagation patterns and incorporates multiple contrastive modules to handle sparsity, zero-shot, and synonymity in OpenKGs.

Findings

TernaryCL outperforms state-of-the-art models on benchmark datasets.

Incorporating ternary propagation improves representation quality.

Contrastive modules effectively handle zero-shot and synonymy issues.

Abstract

Sparsity of formal knowledge and roughness of non-ontological construction make sparsity problem particularly prominent in Open Knowledge Graphs (OpenKGs). Due to sparse links, learning effective representation for few-shot entities becomes difficult. We hypothesize that by introducing negative samples, a contrastive learning (CL) formulation could be beneficial in such scenarios. However, existing CL methods model KG triplets as binary objects of entities ignoring the relation-guided ternary propagation patterns and they are too generic, i.e., they ignore zero-shot, few-shot and synonymity problems that appear in OpenKGs. To address this, we propose TernaryCL, a CL framework based on ternary propagation patterns among head, relation and tail. TernaryCL designs Contrastive Entity and Contrastive Relation to mine ternary discriminative features with both negative entities and relations, introduces Contrastive Self to help zero- and few-shot entities learn discriminative features, Contrastive Synonym to model synonymous entities, and Contrastive Fusion to aggregate graph features from multiple paths. Extensive experiments on benchmarks demonstrate the superiority of TernaryCL over state-of-the-art models.