A Progressive Question Answering Framework Adaptable to Multiple Knowledge Sources

Yirui Zhan, Yanzeng Li, Minhao Zhang, Lei Zou

Abstract: Existing deep learning-based models for knowledge base question answering (KBQA) suffer from the high costs of adapting the system to disparate datasets in real-world scenarios (e.g., multi-tenant platform). In this paper, we present ADMUS, a progressive knowledge base question answering framework designed to accommodate a wide variety of datasets with multiple languages by decoupling the architecture of conventional KBQA systems. Our framework supports the seamless integration of new datasets with minimal effort, only requiring creating a dataset-related micro-service at a negligible cost. To enhance the usability of ADUMS, we design a progressive framework consisting of three stages, ranging from executing exact queries, generating approximate queries and retrieving open-domain knowledge referring from large language models. An online demonstration of ADUMS is available at: https://answer.gstore.cn/pc/index.html.