Multi-Distribution Robust Conformal Prediction was accepted to ICML 2026. Paper Code
Yuqi Yang (Aragorn)
I study mathematical statistics, with a current focus on conformal inference and statistical stepping stones in agentic scientific discovery.
I am pursuing a bachelor's degree at HKUST, majoring in Mathematics and Computer Science. I will start a PhD in Statistics at the Wharton School, University of Pennsylvania, in Fall 2026. I currently work with Prof. Ying Jin at Wharton and Prof. Xinzhou Guo at HKUST. Previously, I worked on cognitive science and intention understanding with Prof. Yangqiu Song at the KnowComp Group, and visited ETH Zurich in Spring 2025.
During my undergrad, I also did internships in quantitative research at Hangzhou Wangdao, hedge fund research at Jihai Investment, and software development at Hundsun.
News & Updates
Research
- Uncertainty and causality. Conformal inference, selective inference, multiple testing, and causal inference.
- Agentic scientific discovery. Automated, rigorous exploration, falsification, and evaluation of AI systems.
Publications
Multi-Distribution Robust Conformal Prediction
[ICML 2026]Summary
One single prediction set that offers valid coverage over multiple heterogeneous distributions.
SessionIntentBench: Multi-task Inter-Session Intention-Shift Modeling Benchmark for E-commerce Customer Behavior Understanding
[ACL 2026 Findings]Summary
Modeling framework for user purchase intent by extracting structured features from e-commerce session interactions.
* Co-first authors.
Selected Awards
- Aug 2024 China Construction Bank Asia Scholarship
- May 2024 S.S. Chern Class Scholarship
- Apr 2024 Chiaphua Industries Limited Scholarship
- Jan 2024 Dean's Honours List
- Oct 2023 University Scholarship Scheme for Continuing UG Student
Misc
Outside of research, I go by Aragorn. It's a nod to my favorite movie series, The Lord of the Rings, and it's the name I've used ever since I first saw the films.