Po-Yi Lu (呂栢頤) I am a fourth-year Ph. D. student at the Graduate Institute of Networking and Multimedia at National Taiwan University. I am also a visiting scholar at Texas A&M University in FLAIR Lab, mentored by Prof. Kuan-Hao Huang and the Associate Learner in CLLab, advised by Prof. Hsuan-Tien Lin. Previously, I worked as a researcher at Chunghwa Telecom Laboratories for five years. My research interests include diffusion models (generative models), preference alignment (AI safety), active learning (human-in-the-loop), and weakly supervised learning (limited-label learning). I also enjoyed promoting human thinking about AI safety and Ethics, specifically by organizing Taiwan AI Safety (TWAIS) Group which is supported by Effective Altruism (EA) community. Email  /  CV  /  Twitter  /  Github  /  Linkedin  /  Google Scholar

News

• [December 2025] I receive the scholarship "2026 Ph.D. Students Study Abroad Program" (115 年國科會千里馬計劃) from NSTC, Taiwan.
• [October 2025] Our paper about re-benchmarking active learning has been awarded a J2C Certification.
• [October 2025] I am visiting Texas A&M University as a visiting scholar for nine months.
• [August 2025] We release TWAIS - Taiwan AI safety workshop (based on ARENA 3.0 courses) with the slides.
• [June 2025] Our paper about re-benchmarking active learning has been accepted by TMLR 2025.
• [May 2025] Our paper about universal domain adaptation has been accepted by ICML 2025.
  
• [February 2025] Our paper about AI labeling for complementary label learning has been accepted by PAKDD 2025.
  
• [January 2025] I will serve as an Assistant Program Chair for NeurIPS 2025.

Research

Conference Publications

Tackling Dimensional Collapse toward Comprehensive Universal Domain Adaptation

Hung-Chieh Fang, Po-Yi Lu, Hsuan-Tien Lin
In Proceedings of the 42nd International Conference on Machine Learning 2025

We investigate the self-supervised loss to tackle dimensional collapse in the target representation and step toward more comprehensive Universal Domain Adaptation.

The Unexplored Potential of Vision-Language Models for Generating Large-Scale Complementary-Label Learning Data

Tan-Ha Mai, Nai-Xuan Ye, Yu-Wei Kuan, Po-Yi Lu, and Hsuan-Tien Lin
In Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD) 2025.

We make Complementary Label Learning practical with Vision-Language Model-based auto-labeling for large-scale data.

Defending Text-to-image Diffusion Models: Surprising Efficacy of Textual Perturbations Against Backdoor Attacks

Oscar Chew, Po-Yi Lu, Jayden Lin, Hsuan-Tien Lin
In the Dark Side of Generative AIs and Beyond Workshop @ ECCV, 2024.

We propose straightforward textual perturbations on prompts to defend against backdoor attacks for text-to-image tasks.

A More Robust Baseline for Active Learning by Injecting Randomness to Uncertainty Sampling

Po-Yi Lu, Chun-Liang Li, and Hsuan-Tien Lin
In the Artificial Intelligence & Human Computer Interaction @ ICML 2023.

We investigate the injecting of slight randomness into uncertainty sampling to balance the bias in the pure uncertainty sampling with a small variance.

Journal Publications

An Expanded Benchmark that Rediscovers and Affirms the Edge of Uncertainty Sampling for Active Learning in Tabular Datasets

Po-Yi Lu, Yi-Jie Cheng, Chun-Liang Li, and Hsuan-Tien Lin
Transactions on Machine Learning Research (TMLR), 2025. Reproducibility Certification   J2C

We re-benchmark active learning in tabular datasets and affirm the effectiveness of uncertainty sampling.

Experience

Education

• September 2013-June 2015, M.S. @ Institute of Industrial Engineering, National Taiwan University
• September 2009-June 2013, B.C. @ Department of Mathematical Science, National ChengChi University

Work

• May 2018-May 2021, Researcher @ Telecommunication Laboratories, Chunghwa Telecom
• December 2015-April 2018, Associate Researcher @ Telecommunication Laboratories, Chunghwa Telecom

Services

Academic Activities

• Program Chair Assistant: NeurIPS 2025
• Session chair: ACML 2024, NeurIPS 2025
• Reviewer: AISTATS 2024, ACML 2024, AISTATS 2025, NeurIPS D&B Track 2025, ICLR 2026, AISTATS 2026

Teaching

• September 2020-December 2023, Instructor @ Information Club, Chingshin Academy
• Fall 2023, Machine Learning, (as Teaching Assistant)
• Spring 2023, Machine Learning, (as Head of Teaching Assistant of 12 TAs)
• Spring 2015, Linear Algebra and Its Applications (IE5034)
• Fall 2014, Introduction to Statistical Control and Optimization (IE5049)

Community

• Organizer of AI Safety Taiwan group

Awards

• 2020 T-Brain E.SUN NLP AML detection competition: Rank 19/409 (5%)
• 2017 Taiwan Innotech Expo, Gold Medal Award
• 2015 Big Data Analytics for Semiconductor Manufacturing by TSMC, Honourable Mention Award

This template is adapted from Jon Barron.