Learning to Infer from Unlabeled Data: A Semi-supervised Learning Approach for Robust Natural Language Inference

TL;DR

This paper introduces a semi-supervised learning framework for Natural Language Inference that leverages a conditional language model to generate hypotheses for unlabeled premises, significantly improving performance in low-resource scenarios.

Contribution

It proposes a novel SSL approach for NLI using BART to generate hypotheses, addressing the challenge of unlabeled sentence pairs in semi-supervised learning.

Findings

Substantial performance improvements on four NLI datasets in low-resource settings.

Effective utilization of unlabeled data through hypothesis generation with BART.

Code release facilitates reproducibility and further research.

Abstract

Natural Language Inference (NLI) or Recognizing Textual Entailment (RTE) aims at predicting the relation between a pair of sentences (premise and hypothesis) as entailment, contradiction or semantic independence. Although deep learning models have shown promising performance for NLI in recent years, they rely on large scale expensive human-annotated datasets. Semi-supervised learning (SSL) is a popular technique for reducing the reliance on human annotation by leveraging unlabeled data for training. However, despite its substantial success on single sentence classification tasks where the challenge in making use of unlabeled data is to assign "good enough" pseudo-labels, for NLI tasks, the nature of unlabeled data is more complex: one of the sentences in the pair (usually the hypothesis) along with the class label are missing from the data and require human annotations, which makes SSL for NLI more challenging. In this paper, we propose a novel way to incorporate unlabeled data in SSL for NLI where we use a conditional language model, BART to generate the hypotheses for the unlabeled sentences (used as premises). Our experiments show that our SSL framework successfully exploits unlabeled data and substantially improves the performance of four NLI datasets in low-resource settings. We release our code at: https://github.com/msadat3/SSL_for_NLI.