NLP for One Health

From BERT to ChatGPT

The use of artificial intelligence (AI) in One Health has the potential to revolutionize disease detection and outbreak response. Natural Language Processing (NLP) is a subfield of AI that enables computers to understand and analyze large amounts of text data, including social media posts, news articles or medical records.

In this practical session, we will explore how NLP can help in the early detection of outbreaks and monitor crisis situations by social mining of newspapers and social media. Specifically, we will start by discussing BERT, a powerful pre-trained language model that can be fine-tuned for specific NLP tasks. We will then move on to ChatGPT, the most popular large language model that can generate human-like responses to natural language inputs.

Through the use of these 2 models, we will discuss practical examples of how NLP can be applied in the One Health context, including case studies of outbreak detection and social media monitoring. Participants will have the opportunity to work with BERT and ChatGPT in hands-on exercises to gain practical experience with these powerful tools. By the end of the session, participants will have a deeper understanding of how NLP can be used in the One Health context and be equipped with the skills to apply these techniques in their own work.

Speakers:

• Rémy DECOUPES - Research engineer UMR TETIS / INRAE
• Maguelonne TEISSEIRE - Prof. UMR TETIS / INRAE

---

Table of contents

Chapter 1: BERT: Google colab link

1.1 Introduction

1.2 NLP tasks with BERT

1.3 Diving into the model representations

1.4 Fine-tuning

1.5 Text generation with BERT

Chapter 2: ChatGPT: Google colab link

2.1 Requirements

2.2 Text generation with ChatGPT

2.3 Langchain

2.4 Hallucination

2.5 Paper-QA

---

Cite this work

@software{Decoupes_Practical_Session_NLP_2023,
    author = {Decoupes, Rémy and Teisseire, Maguelonne},
    doi = {10.5281/zenodo.7817685},
    license = {GNU GPLv3},
    title = {{Practical Session NLP for One Health}},
    url = {https://github.com/tetis-nlp/pratical-session-nlp-for-one-health-murdoch-mood},
    version = {1.0.0},
    year = {2023}
}