Learning Robust Multimodal Representations via Attribute Disentanglement

Haoyue Bai

PhD University: University of Wisconsin--Madison

Graduation Date: August 2026

PhD Advisor: Robert Nowak

Google Scholar: https://scholar.google.com/citations?user=KQ9HkfcAAAAJ

Poster Title: Toward Reliable Visual Intelligence in Open-World Environments

First Author Papers: 2

Requested Mentors Instructions: I understand — my preferred mentors are listed below.

Requested Mentor 1: Kaiming He, [e-mail]

Requested Mentor 2: Tengyu Ma, [e-mail]

Requested Mentor 3: Percy Liang, [e-mail]

Requested Mentor 4: Sergey Levine, [e-mail]

Requested Mentor 5: Phillip Isola, [e-mail]

Requested Mentor 6: Fei-Fei Li, [e-mail]

Requested Mentor 7: Antonio Torralba, [e-mail]

Requested Mentor 8: Trevor Darrell, [e-mail]

Academia Or Industry: Academia

Mentor Questions:

1. How do you develop strong research taste, especially in identifying which problems in computer vision and multimodal learning are likely to remain important over the next 5–10 years, rather than just following current trends?

2. Many researchers today work across multiple areas (e.g., vision, language, theory). How can I build a clear and distinctive research identity while still pursuing interdisciplinary work, without appearing fragmented?

3. My work combines theoretical guarantees with empirical studies of large-scale models. How can I best balance rigor and real-world relevance so that my research is both principled and impactful in practice?

CVPR Participation: I will be presenting a paper at the main conference, I will be attending as a general participant

Previous DC Certification: I certify that I have not previously attended a Doctoral Consortium at ICCV, ECCV or CVPR.

In Person Confirmation: I confirm that I plan to attend CVPR 2026 in person.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 121