MultiMed-ST / README.md
leduckhai's picture
Update README.md
7ce0436 verified
metadata
viewer: true
dataset_info:
  - config_name: Chinese
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: text
        dtype: string
      - name: traditional_chinese
        dtype: string
      - name: English
        dtype: string
      - name: Vietnamese
        dtype: string
      - name: French
        dtype: string
      - name: German
        dtype: string
    splits:
      - name: train
        num_examples: 1242
      - name: eval
        num_examples: 91
      - name: corrected.test
        num_examples: 225
  - config_name: English
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: text
        dtype: string
      - name: Vietnamese
        dtype: string
      - name: Chinese
        dtype: string
      - name: traditional_chinese
        dtype: string
      - name: French
        dtype: string
      - name: German
        dtype: string
      - name: source
        dtype: string
      - name: link
        dtype: string
      - name: type
        dtype: string
      - name: topic
        dtype: string
      - name: icd-10 code
        dtype: string
      - name: speaker
        dtype: string
      - name: role
        dtype: string
      - name: gender
        dtype: string
      - name: accent
        dtype: string
    splits:
      - name: train
        num_examples: 25512
      - name: eval
        num_examples: 2816
      - name: corrected.test
        num_examples: 4751
  - config_name: French
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: text
        dtype: string
      - name: English
        dtype: string
      - name: Vietnamese
        dtype: string
      - name: Chinese
        dtype: string
      - name: traditional_chinese
        dtype: string
      - name: German
        dtype: string
    splits:
      - name: train
        num_examples: 1403
      - name: eval
        num_examples: 42
      - name: corrected.test
        num_examples: 344
  - config_name: German
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: text
        dtype: string
      - name: English
        dtype: string
      - name: Vietnamese
        dtype: string
      - name: Chinese
        dtype: string
      - name: traditional_chinese
        dtype: string
      - name: French
        dtype: string
    splits:
      - name: train
        num_examples: 1443
      - name: eval
        num_examples: 287
      - name: corrected.test
        num_examples: 1091
  - config_name: Vietnamese
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: duration
        dtype: float64
      - name: text
        dtype: string
      - name: English
        dtype: string
      - name: Chinese
        dtype: string
      - name: traditional_chinese
        dtype: string
      - name: French
        dtype: string
      - name: German
        dtype: string
    splits:
      - name: train
        num_examples: 4548
      - name: eval
        num_examples: 1137
      - name: corrected.test
        num_examples: 3437
configs:
  - config_name: Chinese
    data_files:
      - split: train
        path: chinese/train-*
      - split: eval
        path: chinese/eval-*
      - split: corrected.test
        path: chinese/corrected.test-*
  - config_name: English
    data_files:
      - split: train
        path: english/train-*
      - split: eval
        path: english/eval-*
      - split: corrected.test
        path: english/corrected.test-*
  - config_name: French
    data_files:
      - split: train
        path: french/train-*
      - split: eval
        path: french/eval-*
      - split: corrected.test
        path: french/corrected.test-*
  - config_name: German
    data_files:
      - split: train
        path: german/train-*
      - split: eval
        path: german/eval-*
      - split: corrected.test
        path: german/corrected.test-*
  - config_name: Vietnamese
    data_files:
      - split: train
        path: vietnamese/train-*
      - split: eval
        path: vietnamese/eval-*
      - split: corrected.test
        path: vietnamese/corrected.test-*
task_categories:
  - translation
  - automatic-speech-recognition
language:
  - vi
  - en
  - de
  - zh
  - fr
license: mit
tags:
  - medical

MultiMed-ST: Large-scale Many-to-many Multilingual Medical Speech Translation

Preprint

Khai Le-Duc*, Tuyen Tran*,
Bach Phan Tat, Nguyen Kim Hai Bui, Quan Dang, Hung-Phong Tran, Thanh-Thuy Nguyen, Ly Nguyen, Tuan-Minh Phan, Thi Thu Phuong Tran, Chris Ngo,
Nguyen X. Khanh**, Thanh Nguyen-Tang**
*Equal contribution
**Equal supervision
  • Abstract: Multilingual speech translation (ST) in the medical domain enhances patient care by enabling efficient communication across language barriers, alleviating specialized workforce shortages, and facilitating improved diagnosis and treatment, particularly during pandemics. In this work, we present the first systematic study on medical ST, to our best knowledge, by releasing MultiMed-ST, a large-scale ST dataset for the medical domain, spanning all translation directions in five languages: Vietnamese, English, German, French, Traditional Chinese and Simplified Chinese, together with the models. With 290,000 samples, our dataset is the largest medical machine translation (MT) dataset and the largest many-to-many multilingual ST among all domains. Secondly, we present the most extensive analysis study in ST research to date, including: empirical baselines, bilingual-multilingual comparative study, end-to-end vs. cascaded comparative study, task-specific vs. multi-task sequence-to-sequence (seq2seq) comparative study, code-switch analysis, and quantitative-qualitative error analysis. All code, data, and models are available online: https://github.com/leduckhai/MultiMed-ST.

Please press ⭐ button and/or cite papers if you feel helpful.

@article{le2025multimedst,
  title={MultiMed-ST: Large-scale Many-to-many Multilingual Medical Speech Translation},
  author={Le-Duc, Khai and Tran, Tuyen and Tat, Bach Phan and Bui, Nguyen Kim Hai and Dang, Quan and Tran, Hung-Phong and Nguyen, Thanh-Thuy and Nguyen, Ly and Phan, Tuan-Minh and Tran, Thi Thu Phuong and others},
  journal={arXiv preprint arXiv:2504.03546},
  year={2025}
}

Dataset and Models:

Dataset: HuggingFace dataset

Fine-tuned models: HuggingFace models

Contact:

Core developers:

Khai Le-Duc

University of Toronto, Canada
Email: [email protected]
GitHub: https://github.com/leduckhai

Tuyen Tran

Hanoi University of Science and Technology, Vietnam
Email: [email protected]

Bui Nguyen Kim Hai

Eötvös Loránd University, Hungary
Email: [email protected]