Unsupervised Pre-training with Structured Knowledge for Improving Natural Language Inference

TL;DR

This paper explores combining unsupervised pretraining with structured knowledge to enhance natural language inference, demonstrating improved performance over existing models and potential applicability to other sentence classification tasks.

Contribution

It introduces models that integrate structured knowledge into pre-trained models for NLI, effectively combining two knowledge sources for better performance.

Findings

Models outperform previous BERT-based state-of-the-art

Structured knowledge integration improves inference accuracy

Approach can extend to other sentence classification tasks

Abstract

While recent research on natural language inference has considerably benefited from large annotated datasets, the amount of inference-related knowledge (including commonsense) provided in the annotated data is still rather limited. There have been two lines of approaches that can be used to further address the limitation: (1) unsupervised pretraining can leverage knowledge in much larger unstructured text data; (2) structured (often human-curated) knowledge has started to be considered in neural-network-based models for NLI. An immediate question is whether these two approaches complement each other, or how to develop models that can bring together their advantages. In this paper, we propose models that leverage structured knowledge in different components of pre-trained models. Our results show that the proposed models perform better than previous BERT-based state-of-the-art models. Although our models are proposed for NLI, they can be easily extended to other sentence or sentence-pair classification problems.