ZS4IE: A toolkit for Zero-Shot Information Extraction with simple Verbalizations

TL;DR

This paper introduces ZS4IE, a toolkit enabling zero-shot information extraction through user verbalizations of entities and relations, leveraging a Textual Entailment model to reduce annotation effort and improve flexibility.

Contribution

The paper presents a novel zero-shot IE workflow and a user-friendly toolkit that allows direct verbalization of entities and relations, simplifying the annotation process.

Findings

Achieves strong zero-shot IE performance with minimal user effort

Demonstrates effectiveness across four IE tasks

Open-source implementation available for community use

Abstract

The current workflow for Information Extraction (IE) analysts involves the definition of the entities/relations of interest and a training corpus with annotated examples. In this demonstration we introduce a new workflow where the analyst directly verbalizes the entities/relations, which are then used by a Textual Entailment model to perform zero-shot IE. We present the design and implementation of a toolkit with a user interface, as well as experiments on four IE tasks that show that the system achieves very good performance at zero-shot learning using only 5--15 minutes per type of a user's effort. Our demonstration system is open-sourced at https://github.com/BBN-E/ZS4IE . A demonstration video is available at https://vimeo.com/676138340 .