Xumin Liu

Full Professor, Computer Science, Rochester Institute of Technology

Joined

July 2019

Names

Xumin Liu

Emails

****@cs.rit.edu (Confirmed)

****@rit.edu (Confirmed)

Personal Links

Homepage

Google Scholar

DBLP

Career & Education History

Full Professor

Computer Science, Rochester Institute of Technology (rit.edu)

2009 – Present

Advisors, Relations & Conflicts

PhD Advisee

Yuansheng Zhu

Present

PhD Advisee

Spandan Pyakurel

Present

PhD Advisee

Xiaofan Que

Present

PhD Advisee

Mahsa Mozaffari

Present

PhD Advisee

Dingrong Wang

Present

PhD Advisee

Anusha Acharya

2024 – Present

PhD Advisee

Abhinab Acharya

2022 – Present

PhD Advisor

Athman Bouguettaya

2003 – 2009

Expertise

Deep Learning

Machine Learning

2008 – 2024

Natural Language Processing

2008 – 2024

Data Science

2003 – 2024

Publications

Multi-Source Knowledge-Fusion for Source-Free Domain Adaptation in Object Detection
Xiaofan Que, Dingrong Wang, Daniel Krutz, Xumin Liu, Qi Yu
- Submitted to ICLR 2026
- Readers: Everyone
Learning to Compose: Continual Visual QA through a Dual-Purpose Mixture-of-Experts Framework
Md Abdullah Al Forhad, Yuansheng Zhu, Xumin Liu, Qi Yu, Weishi Shi
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
Vulnerability-Aware Parameter-Efficient Fine-Tuning for Enhanced Adversarial Robustness
Abhinab Acharya, Yuansheng Zhu, Dayou Yu, Qi Yu, Xumin Liu
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
Diversity–Ambiguity Exploration for Weakly Supervised Video Anomaly Detection
Anusha Acharya, Hitesh Sapkota, Qi Yu, Xumin Liu
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
Bayesian-Informed Diverse Sampling for Calibration of Fine-Tuned Foundation Models with Evidential Ensembles
Spandan Pyakurel, Hitesh Sapkota, Ervine Zheng, Dingrong Wang, Xumin Liu, Qi Yu
- Submitted to ICLR 2026
- Readers: Everyone
InsCal: Calibrated Multi-Source Fully Test-Time Prompt Tuning for Object Detection
Xiaofan Que, Dingrong Wang, Xumin Liu, Qi Yu
- ICLR 2026 Conference Withdrawn Submission
- Readers: Everyone
Parameter-Efficient Multi-Source Domain-Adaptive Prompt Tuning for Open-Vocabulary Object Detection
Xiaofan Que, Dingrong Wang, Daniel Krutz, Xumin Liu, Qi Yu
- Submitted to ICLR 2026
- Readers: Everyone
Mixing Expertise with Confidence: A Mixture of Expert Framework for Robust Multi-Modal Continual Learner
Md Abdullah Al Forhad, Yuansheng Zhu, Abhinab Acharya, Xumin Liu, Qi Yu, Weishi Shi
- Submitted to ICLR 2026
- Readers: Everyone
Knowledge Exchange with Confidence: Cost-Effective LLM Integration for Reliable and Efficient Visual Question Answering
Mahsa Mozaffari, Hitesh Sapkota, Xumin Liu, Qi Yu
- ICLR 2026 Poster
- Readers: Everyone
Avoid Being a Shortcut Learner through Library-Based Re-Learning
Abhinab Acharya, Dayou Yu, Yuansheng Zhu, Qi Yu, Xumin Liu
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone

View all 86 publications

Co-Authors

View all 109 co-authors