Kaggle Aivle School 4th MiniProject 스팸메일 분류

Official implementation of "Using Pre-Trained Language Models in an End-to-End Pipeline for Antithesis Detection" accepted in LREC-2024

A binary text classifier that uses the ELECTRA model.

Work focus on Transformer model to Start classification (1-5) about reviews of YELP.

Solving Math Word Problems Using Language Models and Contrastive Loss

Code of paper: "ELECTRA is a Zero-shot learner, Too" -- Prompt-based ELECTRA for zero-shot learning.

MSc thesis project: classification of italian certified electronic mails using SOTA machine learning, fine-tuning pre-trained deep learning models and data augmentation techniques

Named entity recognition model with Dialog-ELECTRA

Official implementation of "Using Pre-Trained Language Models in an End-to-End Pipeline for Antithesis Detection" accepted in LREC-2024

Korean Word-Spacing with KoCharELECTRA

Simple NER Pipeline Using KoCharELECTRA

上海交通大学自然语言理解2020春：CoLA Task

ML and Natual Language Processing

Biomedical Question Answering system developed as part of my Masters' Thesis.

Baseline code for Korean open domain question answering(ODQA)

Simple Text Classification[WIP]

Question Answering and Question Generation NLP tasks on the SQuAD v1.1 dataset

Sarcasm is a term that refers to the use of words to mock, irritate, or amuse someone. It is commonly used on social media. The metaphorical and creative nature of sarcasm presents a significant difficulty for sentiment analysis systems based on affective computing. The technique and results of our team, UTNLP, in the SemEval-2022 shared task 6 …

AiSpace: Better practices for deep learning model development and deployment For Tensorflow 2.0

Distilling Task-Specific Knowledge from Teacher Model into BiLSTM