INF582-2023 Challenge: Human Text vs. AI Text

Ecole Polytechnique Fall 2023

Project description Text Classification task.

Custom dataset with text and its associated labels.

TF-IDF + Logistic Regression
Doc2Vec +Logistic Regression
Fine-tuning language models
- Bert with custom classification head
- RoBERTa with custom classification head
- GPT-2 with custom classification head
Transfer learning with language models
Bert with custom classification head
RoBERTa with custom classification head
GPT-2 with custom classification head

Demo_AI_Text_Classification.ipynb files : script for finetuning Bert, RoBERTa and GPT-2
doc_2_vec-model.ipynb : Doc2Vec model
transfer-learning-demo.ipynb : script for transfer learning with BERT,RoBERTa, and GPT-2 train.py : transfer learning with tunable parameters: epochs,max_length, validation split.
submission.py : Script for making inference on the test data set.

Name		Name	Last commit message	Last commit date
__pycache__		__pycache__
config		config
models		models
utils		utils
.gitignore		.gitignore
Demo_AI_Text_Classification.ipynb		Demo_AI_Text_Classification.ipynb
IN582_2023_Challenge.pdf		IN582_2023_Challenge.pdf
Preprocessing .ipynb		Preprocessing .ipynb
README.md		README.md
TF-IDF.ipynb		TF-IDF.ipynb
doc_2_vec-model.ipynb		doc_2_vec-model.ipynb
doc_vec_tools.py		doc_vec_tools.py
submission.py		submission.py
train.py		train.py
transfer-learning-demo.ipynb		transfer-learning-demo.ipynb