I am an undergraduate student in Communication Engineering at Northeastern University. My research interests include LLM agents and retrieval-augmented reasoning, trustworthy multimodal AI, and recommender systems. I am currently seeking PhD opportunities for Fall 2027.
🔥 News
- 2026.08: 🎉🎉 Our paper on open-set standpoint screening was accepted by the EMNLP 2026 Main Conference.
- 2026.08: 🎉🎉 Three papers on uncertainty-aware recommendation, multimodal fake-news detection, and cold-start recommendation were accepted by WISE 2026.
- 2026.05: 🏆 Received the Finalist Award (Top 1%) in the COMAP Mathematical Contest in Modeling.
- 2026.05: 🎉🎉 R3-REC appeared at ICASSP 2026 in Barcelona.
- 2025: 🏆 Received the National Third Prize in the 18th Advanced Robot and Simulation Technology Competition.
- 2023–2026: 🏆 Received the University Third-Class Scholarship five times.
📝 Selected Publications
The “10th Juror”: Open-Set Standpoint Screening for Bureaucratic Bias Detection
Yuchen Miao, Zijun Wang, Chang Han, Yurui Shi, Mingtai Zhang, and Siyang Xu
Anatomy of a Decision: Uncertainty-aware Hierarchical Intent Learning via Flow Matching for Multimodal Recommendation
Yuchen Miao, Zijun Wang, Ke Liu, and Siyang Xu
IMEX-FND: A Traceable Interaction-Aware Mixture-of-Experts Framework for Multimodal Fake News Detection
Yuchen Miao, Zijun Wang, Ke Liu, Peixuan Wang, and Chang Han
R3-REC: Reasoning-Driven Recommendation via Retrieval-Augmented LLMs over Multi-Granular Interest Signals
Yuchen Miao, Mingxuan Cui, Yitong Zhu, Yu Wang, and Siyang Xu
MOTIF: Motivation-guided Topology Inference for Cold-start Multimodal Recommendation
Yurui Shi$^\dagger$, Yuchen Miao$^\dagger$, Ximing Hu, Zijun Wang, and Chang Han
($\dagger$ means equal contribution)
PDF · arXiv · Conference
Preprints
- Universal Refinement without Interaction: Order-Optimal 1-Bit Mean Estimation. Yuchen Miao. 2026. arXiv
- Sharp Root Anti-Concentration via Projective Incidence and Ordered Root Laws. Zijun Wang, Yuchen Miao, Yifan Hu, and Huanmin Liu. 2026. arXiv
💻 Research & Experience
- 2025–Present, Hebei Key Laboratory of Marine Perception Network and Data Processing. Research on explainable multimodal recommendation and fake-content detection using LLM reasoning, retrieval augmentation, graph learning, flow matching, and mixture-of-experts models.
- 2025–Present, National College Students’ Innovation and Entrepreneurship Training Program, Project No. 202619145026. Research on explainable recommendation and fake-content detection for new-media security.
- 2025.06–2025.08, Information Technology Intern, China Development Bank, Shanxi Branch.
- 2024.07–2024.09, Big Data Center Intern, Huajin Coking Coal Co., Ltd.
- 2024.07–2024.10, Embedded Systems Development Intern, Beijing Kedinoyi Technology Co., Ltd. (Outstanding Intern)
Patent
面向多模态虚假内容检测的多域混合专家解释性方法。 发明人：苗煜晨、王雨、王刚、冯鑫、徐思洋。中国发明专利申请，申请号：202511816918.8；实质审查中。
📖 Education
- 2023–2027, Northeastern University, B.Eng. in Communication Engineering, Sydney Smart Technology College.
- IELTS: 7.0.
🎖 Honors and Awards
- 2026.05, Finalist Award (Top 1%), COMAP Mathematical Contest in Modeling.
- 2025, National Third Prize, 18th Advanced Robot and Simulation Technology Competition.
- 2023–2026, University Third-Class Scholarship, awarded five times.
- 2025, Provincial Second Prize, National College Students Market Research and Analysis Competition.
- 2025, Provincial Second Prize, Hebei Provincial College Students “Research Hebei” Social Survey Program.
Last updated: September 2026.