- e-Tickets Web Application using ASP.NET Framework
A web application for purchasing movie tickets online. Built using MVC architecture, Bootstrap, and Azure SQL database. Some functionalities include but are not limited to:
- Creating, editing, and deleting movies/actors/producers/cinemas.
- Registering for a new account and logging in.
- Add tickets to the cart and integrate Paypal API for checking out.
- User authentication and authorization.
(Code available at https://github.com/tonyshu001/eTickets)
- Automated Fact-Checking System for Climate Change Claims
Given a claim related to climate change, first, find relevant evidence (approximately 3 to 10 pieces of evidence) from an extensive knowledge base(approximately over 1.2 million ground-truth). Then based on the relevant evidence retrieved, classify the claim as SUPPORTS/REFUTES/NOT_ENOUGH_INFO/DISPUTED. Some main techniques employed are:
- Transformer-based pre-trained language models: Sentence Transformer, DistilBERT, T5, RoBERTa.
- Key-word-based similarity measurement: TF-IDF, BM25+.
- Pre-processing and post-processing: stop-word removal, self-training, and top-k evidence selection.
- Model performance was ranked 7th among over 400 participants.