Multi-task Learning of Negation and Speculation for Targeted Sentiment Classification

TL;DR

This paper introduces a multi-task learning approach incorporating negation and speculation detection to improve targeted sentiment classification robustness, highlighting current challenges and providing new datasets for evaluation.

Contribution

It proposes a novel multi-task learning framework that leverages syntactic and semantic auxiliary tasks to enhance targeted sentiment analysis, especially under negation and speculation.

Findings

Multi-task models outperform baseline on challenge datasets.

Transfer learning improves performance on negation and speculation samples.

Overall sentiment classification still faces significant challenges.

Abstract

The majority of work in targeted sentiment analysis has concentrated on finding better methods to improve the overall results. Within this paper we show that these models are not robust to linguistic phenomena, specifically negation and speculation. In this paper, we propose a multi-task learning method to incorporate information from syntactic and semantic auxiliary tasks, including negation and speculation scope detection, to create English-language models that are more robust to these phenomena. Further we create two challenge datasets to evaluate model performance on negated and speculative samples. We find that multi-task models and transfer learning via language modelling can improve performance on these challenge datasets, but the overall performances indicate that there is still much room for improvement. We release both the datasets and the source code at https://github.com/jerbarnes/multitask_negation_for_targeted_sentiment.