Improving Complex Knowledge Base Question Answering via Question-to-Action and Question-to-Question Alignment
This work addresses the semantic and structural gap in complex knowledge base question answering, which is an incremental improvement for researchers and practitioners in natural language processing and knowledge base systems.
The paper tackles the difficulty of converting natural language questions into action sequences for complex knowledge base question answering by introducing an alignment-enhanced framework (ALCQA) that uses question-to-action and question-to-question alignment, resulting in a 9.88% improvement in F1 score on the CQA dataset compared to state-of-the-art methods.
Complex knowledge base question answering can be achieved by converting questions into sequences of predefined actions. However, there is a significant semantic and structural gap between natural language and action sequences, which makes this conversion difficult. In this paper, we introduce an alignment-enhanced complex question answering framework, called ALCQA, which mitigates this gap through question-to-action alignment and question-to-question alignment. We train a question rewriting model to align the question and each action, and utilize a pretrained language model to implicitly align the question and KG artifacts. Moreover, considering that similar questions correspond to similar action sequences, we retrieve top-k similar question-answer pairs at the inference stage through question-to-question alignment and propose a novel reward-guided action sequence selection strategy to select from candidate action sequences. We conduct experiments on CQA and WQSP datasets, and the results show that our approach outperforms state-of-the-art methods and obtains a 9.88\% improvements in the F1 metric on CQA dataset. Our source code is available at https://github.com/TTTTTTTTy/ALCQA.