Scope resolution of predicted negation cues: A two-step neural network-based approach

TL;DR

This paper investigates a two-step neural network approach for negation scope resolution, focusing on how cue detection accuracy impacts overall performance and highlighting the need for improved negation detection methods.

Contribution

It evaluates the use of Bidirectional LSTM for cue detection in negation scope resolution and analyzes the impact of inaccurate cues on model robustness.

Findings

Recurrent layer-only models are most robust to cue inaccuracies.

Bidirectional LSTM is not suitable for negation cue detection.

Inaccurate cues significantly reduce scope resolution performance.

Abstract

Neural network-based methods are the state of the art in negation scope resolution. However, they often use the unrealistic assumption that cue information is completely accurate. Even if this assumption holds, there remains a dependency on engineered features from state-of-the-art machine learning methods. The current study adopted a two-step negation resolving apporach to assess whether a Bidirectional Long Short-Term Memory-based method can be used for cue detection as well, and how inaccurate cue predictions would affect the scope resolution performance. Results suggest that this method is not suitable for negation detection. Scope resolution performance is most robust against inaccurate information for models with a recurrent layer only, compared to extensions with a Conditional Random Fields layer or a post-processing algorithm. We advocate for more research into the application of deep learning on negation detection and the effect of imperfect information on scope resolution.