Autoregressive Structured Prediction with Language Models

TL;DR

This paper introduces an autoregressive approach to structured prediction with pretrained language models, modeling structures as sequences of actions to improve learning of dependencies and achieve state-of-the-art results across multiple NLP tasks.

Contribution

It proposes a novel autoregressive method for structured prediction that preserves structural dependencies, surpassing previous flattening techniques.

Findings

Achieved new state-of-the-art on NER, relation extraction, and coreference resolution.

Model effectively captures in-structure dependencies without information loss.

Outperforms traditional flattening approaches in structured prediction tasks.

Abstract

Recent years have seen a paradigm shift in NLP towards using pretrained language models ({PLM}) for a wide range of tasks. However, there are many difficult design decisions to represent structures (e.g. tagged text, coreference chains) in a way such that they can be captured by PLMs. Prior work on structured prediction with PLMs typically flattens the structured output into a sequence, which limits the quality of structural information being learned and leads to inferior performance compared to classic discriminative models. In this work, we describe an approach to model structures as sequences of actions in an autoregressive manner with PLMs, allowing in-structure dependencies to be learned without any loss. Our approach achieves the new state-of-the-art on all the structured prediction tasks we looked at, namely, named entity recognition, end-to-end relation extraction, and coreference resolution.