English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Large language models (LLMs) pre-trained on massive corpora have demonstratedimpressive few-shot learning ability on many NLP tasks. A common practice is torecast the task into a text-to-text format such that generative LLMs of naturallanguage (NL-LLMs) like GPT-3 can be prompted to solve it. However, it isnontrivial to perform information extraction (IE) tasks with NL-LLMs since theoutput of the IE task is usually structured and therefore is hard to beconverted into plain text. In this paper, we propose to recast the structuredoutput in the form of code instead of natural language and utilize generativeLLMs of code (Code-LLMs) such as Codex to perform IE tasks, in particular,named entity recognition and relation extraction. In contrast to NL-LLMs, weshow that Code-LLMs can be well-aligned with these IE tasks by designingcode-style prompts and formulating these IE tasks as code generation tasks.Experiment results on seven benchmarks show that our method consistentlyoutperforms fine-tuning moderate-size pre-trained models specially designed forIE tasks (e.g., UIE) and prompting NL-LLMs under few-shot settings. We furtherconduct a series of in-depth analyses to demonstrate the merits of leveragingCode-LLMs for IE tasks.

CodeIE: Large Code Generation Models are Better Few-Shot Information Extractors