Publications

Conference Papers


Make an Omelette with Breaking Eggs: Zero-Shot Learning for Novel Attribute Synthesis

Published in Advances in Neural Information Processing Systems, 2022

Innovated techniques and developed a groundbreaking decompose-reassemble scheme for automating annotation of attribute labels in zero-shot classification datasets, and accomplished an 89.74% reduction in annotation costs while maintaining or improving performance of attribute labels for zero-shot classification datasets

Recommended citation: @inproceedings{li2022make, title={Make an Omelette with Breaking Eggs: Zero-Shot Learning for Novel Attribute Synthesis}, author={Li, Yu Hsuan and Chao, Tzu-Yin and Huang, Ching-Chun and Chen, Pin-Yu and Chiu, Wei-Chen}, booktitle={Advances in Neural Information Processing Systems}, year={2022} }
Paper Link | Slides Link | Project Page | Code

Projects


Double Q-learning

Published in 2021

Supplemented key assumptions of Double Q-Learning paper with supporting proofs essential to its findings, providing detailed explanations and relevant equations to enhance understanding

Report Link

Fake News Detection

Published in 2021

Devised a method utilizing BERT to extract embeddings from tweets and replies, and trained a classifier for fake news detection, attaining an F1 score of 0.7149 on test data

Report Link | Code

Fish Tracking

Published in 2019

IDeveloped an object detection model thru IoT technology for pet owners’ convenience to remotely locate / monitor their pets for safety

Code