Qinglin Zhu <br /> (祝清麟)

Qinglin Zhu
(祝清麟)

King's College London
伦敦国王学院

👨‍🎓 About Me

I am a 4th-year PhD student at King’s College London. I am fortunate to be supervised by Professor Lin Gui and Professor Yulan He. My current research focuses on Large Language Model Reasoning in Natural Language Processing (NLP).

I completed both my Bachelor’s and Master’s degrees in Computer Science and Technology at Harbin Institute of Technology (Shenzhen), under the guidance of Professor Ruifeng Xu. During my master’s studies, I engaged in research focused on Stance Detection and Argument Mining.

Besides research, I have interned at Tencent Music Entertainment for nine months and Shopee for four months, and worked at Baidu for five months as a Natural Language Processing Algorithm Engineer. I am currently interning at Microsoft Research Asia in Beijing.

⭐ I am actively seeking job opportunities. Please feel free to contact me about relevant openings.

🔥 News

📝 Publications

(* Equal contribution, † Corresponding author)

🤔 LLM Reasoning, Decoding & Code Intelligence

ICML-2026
sym

Pull Requests as a Training Signal for Repo-Level Code Editing
Qinglin Zhu, Tianyu Chen, Shuai Lu, Lei Ji, Runcong Zhao, Murong Ma, Xiangxiang Dai, Yulan He, Lin Gui, Peng Cheng, Yeyun Gong.

  • Introduces Clean-PR, a pipeline that converts noisy PR diffs into structured edit blocks, yielding 2M training samples across 12 languages.
  • Achieves +13.6% on SWE-bench Lite and +12.3% on SWE-bench Verified, demonstrating the value of real-world PRs for repo-level code editing.
NeurIPS-2026
sym

Latent Refinement Decoding: Enhancing Diffusion Language Models by Refining Belief States
Qinglin Zhu, Yizhen Yao, Runcong Zhao, Yanzheng Xiang, Siya Qi, Amrutha Saseendran, Chen Jin, Philip Alexander Teare, Bin Liang, Yulan He, Lin Gui.

  • Proposes Latent Refinement Decoding (LRD), a two-stage framework that tackles information loss and premature commitment in diffusion-based language models via latent refinement and predictive feedback.
  • Enables faster, globally consistent parallel generation as a principled alternative to autoregressive decoding.
ICML-2025 Spotlight
sym

Soft Reasoning: Navigating Solution Spaces in Large Language Models through Controlled Embedding Exploration GitHub
Qinglin Zhu, Runcong Zhao, Hanqi Yan, Yulan He, Yudong Chen, Lin Gui.

  • Proposes an embedding-based search framework that guides LLM generation by optimising the first token’s embedding.
  • Combines embedding perturbation for controlled exploration with Bayesian optimisation via a verifier-guided objective, balancing exploration and exploitation.
EMNLP-2026 Oral
sym

Follow the Latent Roadmap: Navigating Revocable Decoding for Diffusion LLMs with Anchor Tokens
Yizhen Yao*, Qinglin Zhu*, Runcong Zhao, Xiangxiang Dai, Yanzheng Xiang, Yulan He, Lin Gui.

  • Proposes ASRD, a training-free framework that decouples diffusion-LLM decoding context into trusted Anchor Tokens and uncertain candidates to curb error propagation.
  • Anchor-Guided Generation and Anchor-Perturbed Verification deliver up to +6.4% accuracy and 7.2× throughput on math and code benchmarks.
ACL-2024
sym

Mirror: A Multiple-perspective Self-Reflection Method for Knowledge-rich Reasoning
Hanqi Yan*, Qinglin Zhu* , Xinyu Wang, Lin Gui, Yulan He. GitHub

  • Proposes Mirror, enabling LLMs to reflect from multiple perspectives via Navigator–Reasoner cooperation.
  • Encourages both diverse and consistent reasoning to overcome self-reflection traps.

🔍 Retrieval and Memory for LLMs

NeurIPS-2026
sym

Beyond RAG for Agent Memory: Retrieval by Decoupling and Aggregation
Zhanghao Hu*, Qinglin Zhu*, Runcong Zhao, Di Liang, Hanqi Yan, Yulan He, Lin Gui.

  • Proposes xMemory, which decouples agent memories into semantic components and organises them hierarchically.
  • Retrieves via top-down aggregation to capture diverse themes, outperforming standard RAG on long-horizon agent tasks.
AAAI-2026 Oral
sym

Beyond Perplexity: Let the Reader Select Retrieval Summaries via Spectrum Projection Score
Zhanghao Hu, Qinglin Zhu, Siya Qi, Yulan He, Hanqi Yan, Lin Gui.

  • Proposes SPS, a supervision-free metric to assess semantic alignment between retrieved summaries and LLM representations.
  • Introduces xCompress, an inference-time controller that ranks and compresses retrievals to improve generation and clarify retrieval–generation interaction.
ACL-2025
sym

Beyond Prompting: An Efficient Embedding Framework for Open-Domain Question Answering
Zhanghao Hu, Hanqi Yan, Qinglin Zhu†, Zhenyi Shen, Yulan He, Lin Gui.

  • Proposes EmbQA, an embedding-level framework for open-domain QA that optimizes retrieval with unsupervised contrastive learning and improves answer diversity via exploratory embeddings.

📚 Narrative Understanding

ACL-2026 Demo
sym

SymbolicThought: Integrating Language Models and Symbolic Reasoning for Consistent and Interpretable Human Relationship Understanding
Runcong Zhao, Qinglin Zhu*, Hainiu Xu, Bin Liang, Yulan He, Lin Gui.

  • Proposes SymbolicThought, a human-in-the-loop system combining LLM extraction and symbolic reasoning for character relationship understanding.
  • Supports editable relationship graphs, logical constraints, and interactive conflict resolution.
ACL-2024 Findings
sym

Large Language Models Fall Short: Understanding Complex Relationships in Detective Narratives
Runcong Zhao*, Qinglin Zhu* , Hainiu Xu, Jiazheng Li, Yuxiang Zhou, Yulan He, Lin Gui. GitHub

  • Existing datasets for narrative understanding often fail to represent the complexity and uncertainty of relationships in real-life social scenarios.
  • To address this gap, we introduce a new benchmark, Conan, designed for extracting and analysing intricate character relation graphs from detective narratives.
Preprint
sym

PLAYER*: Enhancing LLM-based Multi-Agent Communication and Interaction in Murder Mystery Games
Qinglin Zhu, Runcong Zhao, Jinhua Du, Lin Gui, Yulan He.

  • We propose PLAYER*, a novel framework for Murder Mystery Games (剧本杀) using an anytime sampling-based planner and a questioning-driven search framework.

😆 Sentiment Analysis and Stance Detection

🗣️ Argumentation Mining and Sequence Labeling

🏅 Honors and Awards

  • 2023.03 Outstanding Master’s Graduate & Outstanding Dissertation for Master’s Degree, HIT (3%)
  • 2023.03 Sailvan Times Scholarship (2%), Harbin Institute of Technology
  • 2021.10 National Scholarship (for Graduate Student, 2%)
  • 2021.10 Pacemaker to Merit Student (2%), Harbin Institute of Technology
  • 2021.03 NLPCC-2021 Shared Task: Argument Pair Extraction (Top 1)
  • 2021.02 SemEval-2021 Task 5 Competition: Toxic Spans Detection (Top 1, Team Leader)
  • 2020.06 Outstanding Bachelor’s Graduate & Outstanding Dissertation for Bachelor’s Degree, HIT (3%)
  • 2019.10 National Scholarship (for Undergraduate Student, 2%)

📖 Educations

💻 Works and Internships

  • 2025.05 - Present, Microsoft Research Asia , Beijing, China.(Internship)
  • 2023.04 - 2023.08, Baidu , Shanghai, China.
  • 2022.06 - 2022.09, Shopee , Shanghai, China.(Internship)
  • 2021.08 - 2022.04, Tencent Music Entertainment Group , Shenzhen, China.(Internship)

🫡 Service

  • Reviewer: ICML, NeurIPS, ICLR, ACL, AAAI, KDD, EMNLP, NAACL, NLPCC.

🗺️ Visitor Map

Last Updated:

September 24, 2026