Knowledge Prompting in Pre-trained Language Model for Natural Language Understanding

TL;DR

This paper introduces KP-PLM, a flexible knowledge prompting framework that enhances pre-trained language models with factual knowledge from knowledge bases, improving performance on natural language understanding tasks.

Contribution

The paper proposes a novel knowledge prompting paradigm and a flexible KP-PLM framework that incorporates knowledge via prompts, avoiding complex internal modifications of PLMs.

Findings

KP-PLM outperforms state-of-the-art methods on multiple NLU tasks.

The framework is effective in both full-resource and low-resource settings.

Knowledge prompts improve factual reasoning in PLMs.

Abstract

Knowledge-enhanced Pre-trained Language Model (PLM) has recently received significant attention, which aims to incorporate factual knowledge into PLMs. However, most existing methods modify the internal structures of fixed types of PLMs by stacking complicated modules, and introduce redundant and irrelevant factual knowledge from knowledge bases (KBs). In this paper, to address these problems, we introduce a seminal knowledge prompting paradigm and further propose a knowledge-prompting-based PLM framework KP-PLM. This framework can be flexibly combined with existing mainstream PLMs. Specifically, we first construct a knowledge sub-graph from KBs for each context. Then we design multiple continuous prompts rules and transform the knowledge sub-graph into natural language prompts. To further leverage the factual knowledge from these prompts, we propose two novel knowledge-aware self-supervised tasks including prompt relevance inspection and masked prompt modeling. Extensive experiments on multiple natural language understanding (NLU) tasks show the superiority of KP-PLM over other state-of-the-art methods in both full-resource and low-resource settings.