Multi-label Text Classification with BERT

This project is based on the following paper:

• Tran, H. T., Vo, H. H. P., & Luu, S. T. (2021). Predicting Job Titles from Job Descriptions with Multi-label Text Classification. arXiv preprint arXiv:2112.11052.

Steps to Reproduce Implementation

Please note that this project was implemented on Google Colab and Google Drive, both of which are required for simple reproduction.

1. First step is to download the required data files (two options):

   • To download untranslated Vietnamese data, you can access through this link: https://github.com/sonlam1102/job-prediction-multilabel-vietnamese/blob/main/dataset/full_dataset.zip
     • This option also requires translation of the dataset which can be done by using Translation_API_utility_notebook.ipynb in src
     • In order for the experiments to run, the translated files must be named as: en_labels.csv, en_test.csv, en_train.csv
   • To download already translated data, you access through this google drive link: https://drive.google.com/drive/folders/1yaVh9Eb60TVZllV6I3WCZmWdWxjaczlB?usp=sharing

2. Create a new folder in Google Drive titled MLC

3. Download and add Multi_Label_Classification_BERT.ipynb and src to MLC, along with the three translated data files listed above.

4. Once steps 1 and 2 are complete, you will be free to run the remaining code and experiment with our model.

Notes

• Please note that training time is several hours due to model complexity and dataset size.
• If you prefer not to train your own network, you can download our models weights from this link: https://drive.google.com/file/d/1fgj6oCSNfYisZvnlRTjjXjdxouD7WuSH/view?usp=sharing