Structured Prediction as Translation between Augmented Natural Languages

TL;DR

This paper introduces TANL, a translation-based framework for structured prediction tasks that simplifies modeling by converting tasks into language translation problems, achieving state-of-the-art results across multiple benchmarks.

Contribution

The paper presents a unified translation-based approach for various structured prediction tasks, outperforming task-specific models and enabling multi-task learning with a single architecture.

Findings

Achieves new state-of-the-art on joint entity and relation extraction

Outperforms existing models on relation classification and semantic role labeling

Effective in low-resource settings due to better label semantics utilization

Abstract

We propose a new framework, Translation between Augmented Natural Languages (TANL), to solve many structured prediction language tasks including joint entity and relation extraction, nested named entity recognition, relation classification, semantic role labeling, event extraction, coreference resolution, and dialogue state tracking. Instead of tackling the problem by training task-specific discriminative classifiers, we frame it as a translation task between augmented natural languages, from which the task-relevant information can be easily extracted. Our approach can match or outperform task-specific models on all tasks, and in particular, achieves new state-of-the-art results on joint entity and relation extraction (CoNLL04, ADE, NYT, and ACE2005 datasets), relation classification (FewRel and TACRED), and semantic role labeling (CoNLL-2005 and CoNLL-2012). We accomplish this while using the same architecture and hyperparameters for all tasks and even when training a single model to solve all tasks at the same time (multi-task learning). Finally, we show that our framework can also significantly improve the performance in a low-resource regime, thanks to better use of label semantics.