A Span Selection Model for Semantic Role Labeling
arXiv:1810.02245
Abstract
We present a simple and accurate span-based model for semantic role labeling (SRL). Our model directly takes into account all possible argument spans and scores them for each label. At decoding time, we greedily select higher scoring labeled spans. One advantage of our model is to allow us to design and use span-level features, that are difficult to use in token-based BIO tagging approaches. Experimental results demonstrate that our ensemble model achieves the state-of-the-art results, 87.4 F1 and 87.0 F1 on the CoNLL-2005 and 2012 datasets, respectively.
arXiv commentAccepted by EMNLP 2018
References in corpus (6)
- Exact solutions to the nonlinear dynamics of learning in deep linear neural networks
- Segmental Recurrent Neural Networks
- Frame-Semantic Parsing with Softmax-Margin Segmental RNNs and a Syntactic Scaffold
- Constituency Parsing with a Self-Attentive Encoder
- Close Yet Distinctive Domain Adaptation
- A Minimal Span-Based Neural Constituency Parser