Query-Enhanced Adaptive Semantic Path Reasoning for Inductive Knowledge Graph Completion

TL;DR

The paper addresses inductive knowledge graph completion in the presence of emerging entities by introducing QASPR, which jointly exploits structural and semantic information. It combines a query-dependent masking mechanism to prune noisy edges with a global semantic scoring module to capture long-range dependencies along reasoning paths, enabling robust path-based reasoning. The method employs a Bernoulli-based masking scheme, path-level scoring, and Top-$k$ path embeddings to derive accurate entity representations, optimized via a multi-class log-loss over learned embeddings. Empirical results on WN18RR and FB15k237 variants show state-of-the-art performance, with ablations validating the necessity of both masking and semantic scoring; analyses of the probability multiplier $p_e$ reveal dataset-dependent sensitivity to noise. Overall, QASPR offers a scalable and noise-robust framework for inductive KGC with dynamic, unseen entities in real-world knowledge graphs.

Abstract

Conventional Knowledge graph completion (KGC) methods aim to infer missing information in incomplete Knowledge Graphs (KGs) by leveraging existing information, which struggle to perform effectively in scenarios involving emerging entities. Inductive KGC methods can handle the emerging entities and relations in KGs, offering greater dynamic adaptability. While existing inductive KGC methods have achieved some success, they also face challenges, such as susceptibility to noisy structural information during reasoning and difficulty in capturing long-range dependencies in reasoning paths. To address these challenges, this paper proposes the Query-Enhanced Adaptive Semantic Path Reasoning (QASPR) framework, which simultaneously captures both the structural and semantic information of KGs to enhance the inductive KGC task. Specifically, the proposed QASPR employs a query-dependent masking module to adaptively mask noisy structural information while retaining important information closely related to the targets. Additionally, QASPR introduces a global semantic scoring module that evaluates both the individual contributions and the collective impact of nodes along the reasoning path within KGs. The experimental results demonstrate that QASPR achieves state-of-the-art performance.

Figures (3)

• Figure 1: The motivation of the conventional inductive KGC (left) and the proposed QASPR (right). Conventional inductive KGCs traverse nodes directly over KGs. QASPR sequentially masks the noise structure and captures the long-range semantic dependencies over the whole reasoning path.
• Figure 2: The whole framework of QASPR. Specifically, Query-dependent masking module adaptively masks the noisy structural information, thereby retaining effective structure; Global semantic scoring module captures the long-range dependencies by evaluating the semantics of the whole reasoning path; MLP is utilized to compute the scores for the candidates. m indicates the number of reasoning paths and n denotes number of nodes in every reasoning path.
• Figure 3: The performance with different $p_{e}$ values on WN18RR(V1) and FB15k237(V1).