Chengyi Cai

PhD student, University of Melbourne

Joined

January 2024

Names

Chengyi Cai (Preferred)

Emails

****@unimelb.edu.au (Confirmed)

,

****@student.unimelb.edu.au (Confirmed)

,

****@gmail.com (Confirmed)

Personal Links

Career & Education History

PhD student

University of Melbourne (unimelb.edu.au)

2023 – 2027

Advisors, Relations & Conflicts

PhD Advisor

2023 – 2027

Expertise

Machine Learning

2023 – 2027

Publications

Adapting in the Dark: Efficient and Stable Test-Time Adaptation for Black-Box Models
Yunbei Zhang, Shuaicheng Niu, Chengyi Cai, Feng Liu, Jihun Hamm
- TTU at ICLR 2026 (Main) Oral
- Readers: Everyone
Prime Once, then Reprogram Locally: An Efficient Alternative to Black-Box Model Reprogramming
Yunbei Zhang, Chengyi Cai, Feng Liu, Jihun Hamm
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
Parameter-wise Weighted Model Editing for Efficient and Retentive LLM Unlearning
Chengyi Cai, Zesheng Ye, Jiangchao Yao, Jianzhong Qi, Bo Han, Xiaolu Zhang, Feng Liu, JUN ZHOU
- Submitted to ICLR 2026
- Readers: Everyone
Let's Roll a BiFTA: Bi-refinement for Fine-grained Text-visual Alignment in Vision-Language Models
Yuhao Sun, Chengyi Cai, Jiacheng Zhang, Zesheng Ye, Xingliang YUAN, Feng Liu
- Accepted by TMLR
- Readers: Everyone
Understanding Model Reprogramming for CLIP via Decoupling Visual Prompts
Chengyi Cai, Zesheng Ye, Lei Feng, Jianzhong Qi, Feng Liu
- ICML 2025 poster
- Readers: Everyone
Attribute-based Visual Reprogramming for Vision-Language Models
Chengyi Cai, Zesheng Ye, Lei Feng, Jianzhong Qi, Feng Liu
- ICLR 2025 Poster
- Readers: Everyone
Bayesian-guided Label Mapping for Visual Reprogramming
Chengyi Cai, Zesheng Ye, Lei Feng, Jianzhong Qi, Feng Liu
- NeurIPS 2024 oral
- Readers: Everyone
Sample-specific Masks for Visual Reprogramming-based Prompting
Chengyi Cai, Zesheng Ye, Lei Feng, Jianzhong Qi, Feng Liu
- ICML 2024 Spotlight
- Readers: Everyone

Co-Authors