wikipedia / README.md
MedRAG's picture
Update README.md
d76b9ad verified
metadata
task_categories:
  - question-answering
language:
  - en
tags:
  - medical
  - question answering
  - large language model
  - retrieval-augmented generation
size_categories:
  - 10M<n<100M

The Wikipedia Corpus in MedRAG

This HF dataset contains the chunked snippets from the Wikipedia corpus used in MedRAG. It can be used for medical Retrieval-Augmented Generation (RAG).

News

  • (02/26/2024) The "id" column has been reformatted. A new "wiki_id" column is added.

Dataset Details

Dataset Descriptions

As a large-scale open-source encyclopedia, Wikipedia is frequently used as a corpus in information retrieval tasks. We select Wikipedia as one of the corpora to see if the general domain database can be used to improve the ability of medical QA. We downloaded the processed Wikipedia data from HuggingFace and chunked the text using LangChain as snippets with no more than 1000 characters. This HF dataset contains our ready-to-use chunked snippets for the Wikipedia corpus, including 29,913,202 snippets with an average of 162 tokens.

Dataset Structure

Each row is a snippet of Wikipedia, which includes the following features:

  • id: a unique identifier of the snippet
  • title: the title of the Wikipedia article from which the snippet is collected
  • content: the content of the snippet
  • contents: a concatenation of 'title' and 'content', which will be used by the BM25 retriever

Uses

Direct Use

git clone https://huggingface.co/datasets/MedRAG/wikipedia

Use in MedRAG

>> from src.medrag import MedRAG

>> question = "A lesion causing compression of the facial nerve at the stylomastoid foramen will cause ipsilateral"
>> options = {
    "A": "paralysis of the facial muscles.",
    "B": "paralysis of the facial muscles and loss of taste.",
    "C": "paralysis of the facial muscles, loss of taste and lacrimation.",
    "D": "paralysis of the facial muscles, loss of taste, lacrimation and decreased salivation."
}

>> medrag = MedRAG(llm_name="OpenAI/gpt-3.5-turbo-16k", rag=True, retriever_name="MedCPT", corpus_name="Wikipedia")
>> answer, snippets, scores = medrag.answer(question=question, options=options, k=32) # scores are given by the retrieval system

Citation

@article{xiong2024benchmarking,
    title={Benchmarking Retrieval-Augmented Generation for Medicine}, 
    author={Guangzhi Xiong and Qiao Jin and Zhiyong Lu and Aidong Zhang},
    journal={arXiv preprint arXiv:2402.13178},
    year={2024}
}