Towards Improving Interpretability of Language Model Generation through a Structured Knowledge Discovery Approach
Shuqi Liu, Han Wu, Guanzhi Deng, Jianshu Chen, Xiaoyang Wang, Linqi Song
TL;DR
<3-5 sentence high-level summary>The paper tackles the challenge of interpretability in knowledge-enhanced language generation by introducing a task-agnostic Structured Knowledge Hunter that exploits a two-tier knowledge hierarchy (high-level entities and low-level knowledge triples). It combines a local-global entity-centric encoder with a hierarchical pointer planner to extract coherent knowledge plans that guide language model generation through a template-based conversion, enabling human-readable justification of outputs. Across internal table-to-text (RotoWire-FG) and external knowledge-grounded dialogue (KdConv), the method improves knowledge selection accuracy and text quality, outperforming task-specific baselines and standard LMs while maintaining interpretability. Ablation studies and human evaluation support the benefits of entity-aware encoding, single local-global fusion layer usage, and the multi-task learning losses in preserving hierarchical consistency and factual grounding.
Abstract
Knowledge-enhanced text generation aims to enhance the quality of generated text by utilizing internal or external knowledge sources. While language models have demonstrated impressive capabilities in generating coherent and fluent text, the lack of interpretability presents a substantial obstacle. The limited interpretability of generated text significantly impacts its practical usability, particularly in knowledge-enhanced text generation tasks that necessitate reliability and explainability. Existing methods often employ domain-specific knowledge retrievers that are tailored to specific data characteristics, limiting their generalizability to diverse data types and tasks. To overcome this limitation, we directly leverage the two-tier architecture of structured knowledge, consisting of high-level entities and low-level knowledge triples, to design our task-agnostic structured knowledge hunter. Specifically, we employ a local-global interaction scheme for structured knowledge representation learning and a hierarchical transformer-based pointer network as the backbone for selecting relevant knowledge triples and entities. By combining the strong generative ability of language models with the high faithfulness of the knowledge hunter, our model achieves high interpretability, enabling users to comprehend the model output generation process. Furthermore, we empirically demonstrate the effectiveness of our model in both internal knowledge-enhanced table-to-text generation on the RotoWireFG dataset and external knowledge-enhanced dialogue response generation on the KdConv dataset. Our task-agnostic model outperforms state-of-the-art methods and corresponding language models, setting new standards on the benchmark.