The Case of Imperfect Negation Cues: A Two-Step Approach for Automatic Negation Scope Resolution

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 de-pendency on engineered features from state-of-the-art machine learning methods. The cur-rent study adopted a two-step negation resolving approach 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 the scope resolution performance is most robust against inaccurate information for models with a recurrent layer only, compared to ex-tensions with a conditional random field layer or a post-processing algorithm. We advocate for more research into the application of automated deep learning on negation cue detection and the effect of imperfect information on scope resolution.