About Me
I am currently a second-year Master’s student in Computer Science at Northeastern University, focusing on Retrieval-Augmented Generation. Google Scholar:
Now I am engaged in research internships at NEUIR Lab under the guidance of Associate Professor Zhenghao Liu, as well as at THUNLP, supervised by Yukun Yan.
🔎 Research Interests
My research focuses on agentic systems that integrate external knowledge, efficient reasoning, and scalable infrastructure. These span:
- How can models utilize external knowledge: Retrieval-Augmented Generation (RankCoT, KAIR, Cost-Aware), Deep Research (EigentSearch-Q+), AI Memory.
- How can models reason efficiently: Reasoning Compression (Mixed-Policy Distillation), Multi-Modal Reasoning.
- Agent Infrastructure: UltraRAG 2.0.
👀 I am looking for a Ph.D. position starting in Fall 2027 and would love to explore potential collaborations. Let’s connect!
🔥 News
- 2026.04: 🎉 Our paper EigentSearch-Q+ is accepted by ACM CAIS 2026 Demos!
- 2025.08: 🎉 We released UltraRAG 2.0
, an low-code framework for building complex RAG systems!
- 2025.05: 🎉 Our paper RankCoT is accepted by ACL 2025!
📝 Publications
* indicates equal contribution, and † indicates corresponding author.
When Knowledge Is Not Free: Cost-Aware Evidence Selection in Retrieval-Augmented Generation
Mingyan Wu*, Han Yang*, Omer Ben-Porat, Yftah Ziser†
- This work introduces cost-aware RAG, a setting where retrieved evidence is assigned accesscost tiers and systems must answer under an explicit evidence-access budget. We instantiate this setting by augmenting MS MARCO v2.1 with access-friction tiers and evaluate budgeted evidence selection across generaldomain and domain-specific QA benchmarks.
Reasoning Compression with Mixed-Policy Distillation
Han Yang*, Mingyan Wu*, Bailan He, Zeyu Cao, Sikuan Yan, Kevin Qinghong Lin, Zifeng Ding*†
- This work proposes Mixed-Policy Distillation (MPD), a reasoning compression framework that transfers concise reasoning behavior from a larger-sized teacher to a smaller student by distilling teacher-compressed student trajectories. This preserves student-policy exploration while injecting teacher-guided compression.
EigentSearch-Q+: Enhancing Deep Research Agents with Structured Reasoning Tools
Boer Zhang, Mingyan Wu, Dongzhuoran Zhou, Yuqicheng Zhu, Wendong Fan, Puzhen Zhang, Zifeng Ding, Guohao Li, Yuan He†
- This work introduces Q+, a set of query and evidence processing tools that make web search more deliberate by guiding query planning, monitoring search progress, and extracting evidence from long web snapshots.
RankCoT: Refining Knowledge for Retrieval-Augmented Generation through Ranking Chain-of-Thoughts
Mingyan Wu, Zhenghao Liu†,Yukun Yan†, Xinze Li, Shi Yu, Zheni Zeng, Yu Gu, Ge Yu
- This work leverages the strengths of both ranking and summarization to effectively refine the knowledge from retrieval results, thereby aiding LLMs in generating more accurate responses.
Finding What Matters: Anchoring Context Knowledge with Evolving Indices for Iterative Retrieval
Mingyan Wu*, Zhenghao Liu*†, Xinze Li, Yuqing Lan, Yukun Yan†, Shuo Wang, Cheng Yang, Minghe Yu, Zheni Zeng, Maosong Sun
- This work proposes a Knowledge Anchoring framework for Iterative Retrieval that anchors knowledge within retrieved knowledge to guide LLMs to locate the key information during iterative retrieval and answer generation.
📖 Educations
- 2024.09 - now, M.S. School of Computer Science and Engineering, Northeastern University
- 2020.09 - 2024.07, B.S. School of Computer Science, Yangtze University
💻 Internships
- 2024.04 - now, THUNLP, Tsinghua University, Beijing.
- 2023.10 - now, NEUIR Lab, Northeastern University, Shenyang.
🎼 Amateur
- Dancing 💃: I’ve been danced for about nine years. My favorite dance styles are jazz and hiphop. And I also used to be a part-time dance teacher.
- Guitar 🎸: I am a beginner of guitar and I usually play folk guitar.
-
Swimming 🏊: I’ve recently got into swimming and would like to learn more about it.
If you share these interests, I would be glad to connect and grow together 📞.