QA-for-Event-Extraction Open-source Model - Free Deployment to Boost Zero-shot Event Extraction Tasks

QA For Event Extraction

Developed by veronica320

This model is part of the event extraction system from the ACL2021 paper, based on the RoBERTa-large architecture, fine-tuned using the QAMR dataset for zero-shot event extraction tasks.

Question Answering System

Transformers

#Zero-shot event extraction #Question-answering event recognition #Transfer learning applications

Downloads 21

Release Time : 3/2/2022

Model Overview

By framing event extraction as question-answering queries, it leverages pre-trained QA models to achieve zero-shot event extraction, supporting the identification of specific event elements from text.

Model Features

Zero-shot learning capability

Achieves zero-shot event extraction without requiring labeled data for specific event types through transfer learning.

QA-based event extraction

Transforms event element recognition into a QA task, enhancing the model's generalization ability.

Pre-trained model fine-tuning

Based on the powerful RoBERTa-large pre-trained model, fine-tuned on the QAMR dataset.

Model Capabilities

Event element recognition

QA-based information extraction

Zero-shot transfer learning

Use Cases

News analysis

Violence event detection

Identifies relevant elements of attack events (e.g., victims, locations) from news reports.

Example successfully identified 'person' as the victim.

Intelligence analysis

Security event monitoring

Automatically extracts key security event information from large volumes of text.

🚀 QA-for-Event-Extraction

This is a QA model designed for event extraction, offering solutions for extracting event - related information.

🚀 Quick Start

This QA model is an integral part of the event extraction system, which was introduced in the ACL2021 paper: Zero-shot Event Extraction via Transfer Learning: Challenges and Insights.

✨ Features

Pretrained Architecture: It utilizes the roberta-large architecture.
Fine - tuning Data: The model is fine - tuned on the QAMR dataset.

💻 Usage Examples

Basic Usage

To see how the model works, type a question and a context separated in the right - hand - side textboxs under "Hosted inference API".

Example:

Question: Who was killed?
Context: A car bomb exploded Thursday in a crowded outdoor market in the heart of Jerusalem, killing at least two people, police said.
Answer: people

Advanced Usage

Independent Use: To use the QA model independently, follow the huggingface documentation on AutoModelForQuestionAnswering.
As Part of Event Extraction System: To use it as part of the event extraction system, please check out our Github repo.

📚 Documentation

BibTeX entry and citation info

@inproceedings{lyu-etal-2021-zero,
    title = "Zero-shot Event Extraction via Transfer Learning: {C}hallenges and Insights",
    author = "Lyu, Qing  and
      Zhang, Hongming  and
      Sulem, Elior  and
      Roth, Dan",
    booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers)",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.acl-short.42",
    doi = "10.18653/v1/2021.acl-short.42",
    pages = "322--332",
    abstract = "Event extraction has long been a challenging task, addressed mostly with supervised methods that require expensive annotation and are not extensible to new event ontologies. In this work, we explore the possibility of zero-shot event extraction by formulating it as a set of Textual Entailment (TE) and/or Question Answering (QA) queries (e.g. {``}A city was attacked{''} entails {``}There is an attack{''}), exploiting pretrained TE/QA models for direct transfer. On ACE-2005 and ERE, our system achieves acceptable results, yet there is still a large gap from supervised approaches, showing that current QA and TE technologies fail in transferring to a different domain. To investigate the reasons behind the gap, we analyze the remaining key challenges, their respective impact, and possible improvement directions.",
}

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご