Authors: Chloe Qinyu Zhu, Rickard Stureborg, Bhuwan Dhingra

Published on: February 01, 2024

Impact Score: 8.35

Arxiv code: Arxiv:2402.01783

Summary

• What is new: Use of large language models (LLMs) in a zero-shot setting for detecting vaccine concerns in online discussions, without needing large, expensive training datasets.
• Why this is important: Rapid shifts in vaccine concerns and misinformation, challenging for public health efforts to address effectively.
• What the research proposes: Employing LLMs to identify vaccine concerns in online discourse using cost-effective strategies and zero-shot learning.
• Results: GPT-4 outperforms crowdworker accuracy in identifying vaccine concerns, achieving an F1 score of 78.7%.

Technical Details

Technological frameworks used: nan

Models used: Large language models

Data used: VaxConcerns dataset

Potential Impact

Healthcare, public health organizations, social media platforms

Want to implement this idea in a business?

We have generated a startup concept here: VaxTrend Analytics.