Resolving the Scope of Speculation and Negation using Transformer-Based Architectures

9 Jan 2020  ·  Benita Kathleen Britto, Aditya Khandelwal ·

Speculation is a naturally occurring phenomena in textual data, forming an integral component of many systems, especially in the biomedical information retrieval domain. Previous work addressing cue detection and scope resolution (the two subtasks of speculation detection) have ranged from rule-based systems to deep learning-based approaches. In this paper, we apply three popular transformer-based architectures, BERT, XLNet and RoBERTa to this task, on two publicly available datasets, BioScope Corpus and SFU Review Corpus, reporting substantial improvements over previously reported results (by at least 0.29 F1 points on cue detection and 4.27 F1 points on scope resolution). We also experiment with joint training of the model on multiple datasets, which outperforms the single dataset training approach by a good margin. We observe that XLNet consistently outperforms BERT and RoBERTa, contrary to results on other benchmark datasets. To confirm this observation, we apply XLNet and RoBERTa to negation detection and scope resolution, reporting state-of-the-art results on negation scope resolution for the BioScope Corpus (increase of 3.16 F1 points on the BioScope Full Papers, 0.06 F1 points on the BioScope Abstracts) and the SFU Review Corpus (increase of 0.3 F1 points).

PDF Abstract

Datasets


Results from the Paper


 Ranked #1 on Speculation Scope Resolution on SFU Review Corpus (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Benchmark
Speculation Scope Resolution BioScope : Abstracts XLNet F1 97.87 # 2
Negation Scope Resolution BioScope : Abstracts XLNet F1 95.74 # 2
Speculation Scope Resolution BioScope : Full Papers XLNet F1 96.91 # 1
Negation Scope Resolution BioScope : Full Papers XLNet F1 94.4 # 1
Negation Scope Resolution *sem 2012 Shared Task: Sherlock Dataset RoBERTa F1 91.59 # 3
Speculation Scope Resolution SFU Review Corpus XLNet F1 91.00 # 1
Negation Scope Resolution SFU Review Corpus XLNet F1 91.25 # 1

Methods