Zero-Shot Information Extraction as a Unified Text-to-Triple Translation

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1225–1238 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 1225 Zero-Shot Information Extraction as a Unified Text-to-Triple Translation Chenguang Wang∗, Xiao Liu¶, Zui Chen¶, Haoyun Hong¶, Jie Tang¶, Dawn Song∗ ∗UC Berkeley, ¶Tsinghua University {chenguangwang,dawnsong}@berkeley.edu, jietang@tsinghua.edu.cn {liuxiao21,chenzui19,honghy17}@mails.tsinghua.edu.cn Abstract We cast a suite of information extraction tasks into a text-to-triple translation framework. In- stead of solving each task relying on task- specific datasets and models, we formalize the task as a translation between task-specific in- put text and output triples. By taking the task- specific input, we enable a task-agnostic trans- lation by leveraging the latent knowledge that a pre-trained language model has about the task. We further demonstrate that a simple pre- training task of predicting which relational in- formation corresponds to which input text is an effective way to produce task-specific out- puts. This enables the zero-shot transfer of our framework to downstream tasks. We study the zero-shot performance of this framework on open information extraction (OIE2016, NYT, WEB, PENN), relation classification (FewRel and TACRED), and factual probe (Google-RE and T-REx). The model transfers non-trivially to most tasks and is often competitive with a fully supervised method without the need for any task-specific training. For instance, we sig- nificantly outperform the F1 score of the su- pervised open information extraction without needing to use its training set.1 1 Introduction Information extraction refers to the task of auto- matically extracting structured information from unstructured resources, benefiting a wide range of applications such as information retrieval and knowledge base population. Information extraction covers a great variety of tasks in natural language processing (NLP), such as open information ex- traction and relation classification. For example, given a sentence “Born in Glasgow, Fisher is a graduate of the London Opera Centre”, open infor- mation extraction seeks to extract (Fisher; Born in; Glasgow), and “city_of_birth” is predicted as the 1The code and datasets are available at https:// github.com/cgraywang/deepex. relation between a given pair of entities “Fisher” and “Glasgow” for relation classification. Most current approaches design task-specific pipelines for different information extraction tasks. Yet, this presents two limitations for information ex- traction. First, since most of the approaches employ a task-specific model, it is difficult to leverage a sin- gle pipeline to solve many tasks or adapt a model trained on one task to another without changing any task-specific modules. Second, those super- vised state-of-the-arts are trained on task-specific corpora to predict from a fixed set of task-specific categories, which restricts their usability since ad- ditional labeled data is needed to specify any other classes. Such task-specific labeled data is scarce in information extraction. For example, the largest training set for open information extraction con- tains only 3,200 sentences (Stanovsky et al., 2018). Motivated by this, we aim to solve information extraction tasks within the same framework in a task-agnostic setting. In this paper, we propose a unified framework for information extraction. The basic idea is to treat every information extraction problem …

Content truncated to 500 words