A Sequence-to-Sequence Model for Semantic Role Labeling

TL;DR

This paper introduces a sequence-to-sequence model with attention and copying mechanisms for Semantic Role Labeling, demonstrating its potential and limitations in English SRL tasks.

Contribution

It presents a novel generative approach to SRL using sequence-to-sequence modeling with attention and copying, a first step towards more advanced generative SRL methods.

Findings

Model successfully performs SRL argument labeling on English data.

The copying mechanism improves faithful input regeneration.

Structural decoding constraints are needed for competitive performance.

Abstract

We explore a novel approach for Semantic Role Labeling (SRL) by casting it as a sequence-to-sequence process. We employ an attention-based model enriched with a copying mechanism to ensure faithful regeneration of the input sequence, while enabling interleaved generation of argument role labels. Here, we apply this model in a monolingual setting, performing PropBank SRL on English language data. The constrained sequence generation set-up enforced with the copying mechanism allows us to analyze the performance and special properties of the model on manually labeled data and benchmarking against state-of-the-art sequence labeling models. We show that our model is able to solve the SRL argument labeling task on English data, yet further structural decoding constraints will need to be added to make the model truly competitive. Our work represents a first step towards more advanced, generative SRL labeling setups.