Semantic Role Labeling with Iterative Structure Refinement
arXiv:1909.03285
Abstract
Modern state-of-the-art Semantic Role Labeling (SRL) methods rely on expressive sentence encoders (e.g., multi-layer LSTMs) but tend to model only local (if any) interactions between individual argument labeling decisions. This contrasts with earlier work and also with the intuition that the labels of individual arguments are strongly interdependent. We model interactions between argument labeling decisions through {\it iterative refinement}. Starting with an output produced by a factorized model, we iteratively refine it using a refinement network. Instead of modeling arbitrary interactions among roles and words, we encode prior knowledge about the SRL problem by designing a restricted network architecture capturing non-local interactions. This modeling choice prevents overfitting and results in an effective model, outperforming strong factorized baseline models on all 7 CoNLL-2009 languages, and achieving state-of-the-art results on 5 of them, including English.
References in corpus (9)
- Categorical Reparameterization with Gumbel-Softmax
- Training Very Deep Networks
- The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
- Achieving Human Parity on Automatic Chinese to English News Translation
- Learning Approximate Inference Networks for Structured Prediction
- Iterative Refinement for Machine Translation
- Dependency or Span, End-to-End Uniform Semantic Role Labeling
- Image Segmentation by Iterative Inference from Conditional Score Estimation
- Learning with Fenchel-Young Losses