Text-to-SPARQL Goes Beyond English: Multilingual Question Answering Over Knowledge Graphs through Human-Inspired Reasoning

2025-07-22

AI Summary

To bridge the semantic gap and ensure cross-lingual interpretability in multilingual knowledge graph question answering (KGQA)—specifically in natural language-to-SPARQL translation—this paper proposes mKGQAgent, a modular, interpretable framework. It decomposes multilingual semantic parsing into four distinct, explainable components: query planning, multilingual entity linking, SPARQL template generation, and context-aware optimization. To enhance generalization for low-resource languages, we introduce an experience-pool-based in-context learning mechanism. Furthermore, the framework adopts an LLM-coordinated agent architecture to enable end-to-end multilingual reasoning. Evaluated on the Text2SPARQL 2025 Challenge, mKGQAgent achieves first place on both the DBpedia and enterprise-level multilingual benchmarks, demonstrating substantial improvements in SPARQL query accuracy and cross-lingual robustness.

Technology Category

• Natural Language Processing: Question Answering
• Knowledge Representation and Reasoning: Knowledge Representation Languages
• Multiagent Systems: Multiagent Planning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

Accessing knowledge via multilingual natural-language interfaces is one of the emerging challenges in the field of information retrieval and related ones. Structured knowledge stored in knowledge graphs can be queried via a specific query language (e.g., SPARQL). Therefore, one needs to transform natural-language input into a query to fulfill an information need. Prior approaches mostly focused on combining components (e.g., rule-based or neural-based) that solve downstream tasks and come up with an answer at the end. We introduce mKGQAgent, a human-inspired framework that breaks down the task of converting natural language questions into SPARQL queries into modular, interpretable subtasks. By leveraging a coordinated LLM agent workflow for planning, entity linking, and query refinement - guided by an experience pool for in-context learning - mKGQAgent efficiently handles multilingual KGQA. Evaluated on the DBpedia- and Corporate-based KGQA benchmarks within the Text2SPARQL challenge 2025, our approach took first place among the other participants. This work opens new avenues for developing human-like reasoning systems in multilingual semantic parsing.

Problem

• Multilingual natural-language interfaces for knowledge access
• Transforming natural-language questions into SPARQL queries
• Human-inspired reasoning for multilingual KGQA tasks

Innovation

• Human-inspired modular framework for multilingual KGQA
• LLM agent workflow for query planning and refinement
• Experience pool-guided in-context learning