Open-Domain Frame Semantic Parsing Using Transformers

TL;DR

This paper demonstrates that transformer-based generative models significantly improve the accuracy of open-domain frame semantic parsing, outperforming previous state-of-the-art methods on multiple benchmarks.

Contribution

It introduces a multi-task, transformer-based generative approach that advances the state of the art in frame semantic parsing and PropBank SRL parsing.

Findings

Purely generative encoder-decoder models outperform previous methods.

Mixed decoding multi-task models achieve even higher accuracy.

Multi-task transformer models outperform recent state-of-the-art systems.

Abstract

Frame semantic parsing is a complex problem which includes multiple underlying subtasks. Recent approaches have employed joint learning of subtasks (such as predicate and argument detection), and multi-task learning of related tasks (such as syntactic and semantic parsing). In this paper, we explore multi-task learning of all subtasks with transformer-based models. We show that a purely generative encoder-decoder architecture handily beats the previous state of the art in FrameNet 1.7 parsing, and that a mixed decoding multi-task approach achieves even better performance. Finally, we show that the multi-task model also outperforms recent state of the art systems for PropBank SRL parsing on the CoNLL 2012 benchmark.