CMPUT-622-Project

Team

Student name	CCID
Sohyun Park	sohyun2
Seeratpal Jaura	seeratpa

Repository management

• The repository contain the Data
• The Baseline code, DP BERT model and Keyword mask DP BERT model can be found in the same directory as ReadMe.

Reference of Dataset

• The Coronavirus (COVID-19) geotagged tweets dataset is available at https://ieee-dataport.org/open-access/coronavirus-covid-19-geo-tagged-tweets-dataset Rabindra Lamsal. Coronavirus (covid-19) geo-tagged tweets dataset, 2020. URL https: //dx.doi.org/10.21227/fpsb-jz61.

The dataset is present in Data

Resources used in task implementation

• To understand the Opacus library use with the BERT model. The code implementation is used in the DP BERT model and Keyword mask BERT model

  https://opacus.ai/tutorials/building_text_classifier

• Tutorial followed for the implementation of BERT model for sentiment analysis task

  Chris Tran, Skim AI’s Machine Learning Researcher - https://skimai.com/fine-tuning-bert-for-sentiment-analysis/

Run instruction

• The notebooks are Jupyter notebook, so it could be run on the Google Colab by cloning the github repository. The dependencies are installed in the Jupyter notebook itself. So, you would require to run those cell in jupyter notebook to install dependencies.

• The dataset file should be uploaded from your local machine. Therefore, please download and choose the file when you run the cell for downloading the dataset.

• You should use a GPU session of colab to run it, or you will consume lots of time.