Search papers, labs, and topics across Lattice.
This paper introduces SABET-QA, a novel framework for Question Answering over Temporal Knowledge Graphs (TKGQA) that addresses the limitations of existing embedding-based methods in handling multi-step queries. By employing a bidirectional entity-temporal scoring mechanism and a slot-aware contextualization module, SABET-QA iteratively refines reasoning states and aligns question semantics with temporal embeddings. Experimental results on multiple datasets, including CronQuestions and TimeQuestions, show significant performance improvements over strong baselines, particularly for complex queries involving multiple temporal steps.
SABET-QA outperforms existing methods by effectively refining reasoning states across multiple hops, making it a game-changer for complex temporal queries.
Question Answering over Temporal Knowledge Graphs (TKGQA) requires reasoning over time-sensitive facts, yet existing embedding-based methods struggle with multi-step queries due to single-pass reasoning pipelines. We propose SABET-QA, a framework that iteratively refines reasoning states across multiple hops via a bidirectional entity-temporal scoring mechanism and a slot-aware contextualization module that aligns question semantics with temporal KG embeddings. A differentiable working memory enables progressive hypothesis refinement, while auxiliary temporal boundaries serve as coarse supervision when available. Experiments on CronQuestions, Complex-CronQuestions, MultiTQ, and TimeQuestions demonstrate consistent improvements over strong baselines, particularly on complex multi-step temporal queries.
