Yuanjie Shi
PhD Candidate in Computer Science at Washington State University
I am a PhD candidate in Computer Science at the School of Electrical Engineering and Computer Science, Washington State University, advised by Prof. Yan Yan. I expect to graduate in 2026.
My research develops reliable machine learning methods with statistically valid and distribution-free risk control. I work at the intersection of robust machine learning, uncertainty quantification, conformal prediction, federated learning, and reliable and controllable language models.
Research directions
- Distribution-free uncertainty quantification and risk control: conformal methods for reliable and efficient human–ML collaboration.
- Federated learning under decentralized heterogeneity: uncertainty-aware and privacy-preserving learning under cross-client distribution shift.
- Reliable and controllable language models: uncertainty-aware inference, preference alignment, and reliability analysis for multi-step LLM systems.
My work has appeared at ICML, ICLR, NeurIPS, AISTATS, AAAI, and UAI. I received the WSU EECS Outstanding Dissertation award (2026), the WSU EECS Outstanding Research Assistant award (2025), an ICML Gold Reviewer recognition (2026), and a NeurIPS Top Reviewer recognition (2025).
selected publications
- UAI
Valid and Efficient Uncertainty Quantification for Federated Joint ShiftIn Proceedings of the Forty-Second Conference on Uncertainty in Artificial Intelligence, 2026 - ICML
Inference-Time Conformal Reasoning with Valid Factuality Control for Large Language ModelsIn Proceedings of the Forty-Third International Conference on Machine LearningTian Wang and Yuanjie Shi contributed equally. , 2026 - ICLR
Keep the Best, Forget the Rest: Reliable Alignment with Order-Aware Preference OptimizationIn Proceedings of the Fourteenth International Conference on Learning RepresentationsJiacheng Zhu and Yuanjie Shi contributed equally. , 2026 - AISTATS
Conformal Margin Risk Minimization: An Envelope Framework for Robust Learning under Label NoiseIn Proceedings of the 29th International Conference on Artificial Intelligence and StatisticsYuanjie Shi and Peihong Li contributed equally. , 2026 - ICML
Direct Prediction Set Minimization via Bilevel Conformal Classifier TrainingIn Proceedings of the Forty-Second International Conference on Machine LearningYuanjie Shi and Hamed Shahrokhi contributed equally. , 2025 - NeurIPSConformal Prediction for Class-wise Coverage via Augmented Label Rank CalibrationIn Advances in Neural Information Processing Systems, 2024