I am a first-year Ph.D. student in Computer Science at the University of Ottawa, EECS, supervised by Prof. Changjian Shui. Previously, I completed my M.Sc. in Electrical Engineering at McGill University and Mila — Quebec AI Institute under the supervision of Prof. Tal Arbel. I also received my B.Eng. in Computer Engineering from McGill University.
My research studies predictable AI, with an emphasis on developing theoretically grounded quantities that allow us to anticipate the reliability and consequences of AI systems before their outcomes are observed. My earlier work focused on predicting inference reliability through confidence calibration, my current work extends this perspective to the training process, asking whether the effect of data interventions on population generalization can itself be predicted before the intervention is performed.
📝 Selected Publications
BayesAgent: Bayesian Agentic Reasoning Under Uncertainty via Verbalized Probabilistic Graphical Modeling
Hengguan Huang*, Xing Shen*, Guang-Yuan Hao, Songtao Wang, Lingfa Meng, Dianbo Liu, David Alejandro Duchene, Hao Wang, and Samir Bhatt
*Equal contribution [Exposing and Mitigating Calibration Biases and Demographic Unfairness in MLLM Few-Shot In-Context Learning for Medical Image Classification
Xing Shen, Justin Szeto, Mingyang Li, Hengguan Huang, and Tal Arbel
Improving Robustness and Reliability in Medical Image Classification with Latent-Guided Diffusion and Nested-Ensembles
Xing Shen, Hengguan Huang, Brennan Nichyporuk, and Tal Arbel
IEEE Transactions on Medical Imaging, 2025 
Prompt4Trust: A Reinforcement Learning Prompt Augmentation Framework for Clinically-Aligned Confidence Calibration in Multimodal Large Language Models
Anita Kriz*, Elizabeth Laura Janes*, Xing Shen*, and Tal Arbel
ICCV 2025 Workshops *Equal contribution [Paper] [arXiv] [Code] [Hugging Face]
💡 Writing and Ideas
- 2026.08: Two Targets Behind One RCT Estimate The difference in means from a randomized trial is a single number, but it can be aimed at two different quantities: the average effect among the people who enrolled, or the average effect in the population we actually care about. The point estimate does not change. What changes is which error you are entitled to bound, and randomization only earns you one of the two.
- 2026.08: Provable AI, Predictable AI, and Guardrails A proof is a conditional promise about inputs chosen in advance, while predictability is a forecast about the inputs a system actually meets. Neither property contains the other. A two-term bound makes visible exactly what a certificate leaves on the table, and guardrails turn out to attack only one of those two terms.
📰 News
- 2025.11: Our paper “BayesAgent: Bayesian Agentic Reasoning Under Uncertainty via Verbalized Probabilistic Graphical Modeling” has been accepted to AAAI 2026.
- 2025.10: Our Calibration Guidance Prompt Generator 1.5B model is now live on Hugging Face 🤗 ! [Hugging Face]
🎖 Honors and Awards
- University of Ottawa Ph.D. Research Fellowship (2026–2029)
- Graduate Research and Enhancement Travel (GREAT) Award (2025)
- Healthy Brains, Healthy Lives (HBHL) Graduate Student Fellowship - funded via Canada First Research Excellence Fund (CFREF) (2024–2025)
- McGill University Master’s Research Fellowship (2024–2025)
- McGill Faculty of Engineering Class 2023 Distinction
- McGill Summer Undergraduate Research in Engineering (SURE) Award (2023)
🎓 Education
- 2026.01 – Present: Ph.D. in Computer Science, University of Ottawa, CGPA 10.0/10.0
- 2024.01 – 2026.05: M.Sc. in Electrical Engineering, McGill University and Mila — Quebec AI Institute, CGPA 4.0/4.0
- 2018.09 – 2023.05: B.Eng. in Computer Engineering, McGill University, Graduated with Distinction
💻 Professional Service
- Conference Reviewer: NeurIPS (2025–2026), AISTATS (2025–2026), AAAI (2027)
- Journal Reviewer: IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), Transactions on Machine Learning Research (TMLR), IEEE Journal of Biomedical and Health Informatics (JBHI), Frontiers in Artificial Intelligence, Frontiers in Radiology, Journal of Imaging Informatics in Medicine
- Teaching Assistant: CSI 4106 Introduction to Artificial Intelligence, CSI 3131 Operating Systems, University of Ottawa