About
Machine Learning Researcher / Engineer
- Phone: +886 988 185 871
- City: Taipei, Taiwan
- Degree: Master
- Email: r09922104@ntu.edu.tw
I am currently a visiting student in the Vision Learning Lab (VLLab) at UC Merced, working with Chieh Hubert Lin and Prof. Ming-Hsuan Yang. I also work as a research assistant in the Vision Learning Lab (VLLab) at National Taiwan University EECS, advised by Prof. Yu-Chiang Frank Wang. My Primary research interests lie in Computer Vision, with a particular focus on multi-modal models' adaptability, continual learning, and open-domain 3D articulation synthesis.
Prior to my role at VLLab, I am a Machine Learning Engineer at Dcard, where I designed a recommendation system using machine learning algorithms for over 15 million users.
I completed my bachelor's and master's degrees in Computer Science from National Taiwan University, where I conducted research in the Computational Learning Lab under the supervision of Prof. Hsuan-Tien Lin.
Publication
Towards Affordance-Aware Articulation Synthesis for Rigged Objects
Yu-Chu Yu,
Chieh Hubert Lin,
Hsin-Ying Lee,
Chaoyang Wang,
Yu-Chiang Frank Wang,
Ming-Hsuan Yang
arXiv preprint, December 2024.
ReXTime: A Benchmark Suite for Reasoning-Across-Time in Videos
Jr-Jen Chen,
Yu-Chien Liao,
Hsi-Che Lin,
Yu-Chu Yu,
Yen-Chun Chen,
Yu-Chiang Frank Wang
Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track, December 2024.
Select and Distill: Selective Dual-Teacher Knowledge Transfer for Continual Learning on Vision-Language Models
Yu-Chu Yu,
Chi-Pin Huang,
Jr-Jen Chen,
Kai-Po Chang,
Yung-Hsuan Lai,
Fu-En Yang,
Yu-Chiang Frank Wang
European Conference on Computer Vision (ECCV), September 2024.
Semi-Supervised Domain Adaptation with Source Label Adaptation
In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2023.
[ Paper Link Code]
Resume
Education
National Taiwan University
2020/09 - 2023/01
Master of Computer Science & information Engineering
- GPA: 4.1 / 4.3
- Master Thesis: Semi-Supervised Domain Adaptation with Source Label Adaptation (CVPR'23) [Link]
National Taiwan University
2016/09 - 2020/06
Bachelor of Computer Science & information Engineering
Teaching
Data Structures & Algorithms CSIE 1212
2020 Spring
Teaching Assistant. Instructor: Hsuan-Tien Lin
- Assist in the administration of a 400-person class to ensure smooth operation.
- Provide office hours for debugging and answering student questions.
- Collaborate with the instructor to create and grade assignments / exams.
Machine Learning CSIE 5043
2020 Fall
2021 Fall
Teaching Assistant. Instructor: Hsuan-Tien Lin
- Assist in the administration of a 600-person class to ensure smooth operation.
- Provide office hours for answering student questions, helping students understand foundational concepts of machine learning.
- Design and create assignments / projects that allowed students to apply their knowledge in both practical and theoretical settings.
Experience
Vision Learning Lab University of California, Merced
2024/04 - Present
Visiting Student Researcher. Advisor: Ming-Hsuan Yang
- Research on 3D articulated objects synthesis with affordance.
- Develop an affordance-aware pipeline to synthesize 3D articulated objects.
Vision Learning Lab National Taiwan University
2023/09 - Present
Research Assistant. Advisor: Yu-Chiang Frank Wang
- Research on Continual Learning for Vision-Language Models, develop a selective knoweledge distillation algorithm to preserve both previous knowledge and zero-shot transferability for VLMs. (ECCV'24) [Link]
Dcard Taiwan Ltd. Taipei, TW
2022/11 - 2023/09
Machine Learning Engineer
- Design a recommendation system for a Taiwanese forum platform, serving over 15 million users with more than 10,000 posts per day.
- Develop an interactive API to recommend posts to users on-the-fly (< 1.8 sec. for 95% requests) according to their previous behaviours.
Microsoft Taipei, TW
2021/12 - 2022/06
Research Development Intern @ MS Maps & Geospatial Team
- Develop an asynchronous web crawler system to automatically evaluate the performance of reverse geocoding services provided by google, bing and tomtom.
Computational Learning Lab National Taiwan University
2020/04 - 2023/01
Graduate Research Student. Advisor: Hsuan-Tien Lin
- Research on Semi-Supervised Domain Adaptation, introduce a novel source-adaptive paradigm by addressing Domain Adaptation as a Noisy Label Learning problem. Develop a dynamic algorithm to learn from the noisy source labels and outperform state-of-the-art algorithms. (CVPR'23) [Link]
