Zhijiang Guo

Zhijiang Guo is an Assistant Professor at the DSA Thrust, HKUST (GZ), and an Affiliated Assistant Professor of HKUST. Previously, he was a Senior Researcher at Huawei Noah's Ark Lab and a Postdoc at the Department of Computer Science and Technology at the University of Cambridge, where he was also a member of Trinity College.

Dr. Guo earned his PhD in Computer Science from Singapore University of Technology and Design (SUTD) in 2020 under Professor Wei Lu. During his doctoral studies, he was a visiting student at the University of Edinburgh (2019-2020), collaborating with Professors Shay Cohen and Giorgio Satta on Structured Prediction. His undergraduate education was completed at Sun Yat-sen University.

Dr. Guo's research focuses on natural language processing and machine learning, with particular emphasis on large language models (LLMs). His work explores fundamental questions about knowledge representation and reasoning capabilities in LLMs, examining how these systems understand and process information. Key research areas include logical consistency in language models, efficient code generation, autoformalization techniques, meta-reasoning benchmarks, and parameter-efficient fine-tuning methods. His work bridges theoretical insights with practical applications requiring systematic reasoning and knowledge representation.

Dr. Guo's publication record demonstrates significant contributions to advancing reasoning capabilities in language models. His work spans top-tier conferences including ICML (with spotlight papers in 2025), NeurIPS (including an oral presentation in 2024), ICLR (with spotlight papers in 2024), and ACL. His research shows a clear progression from foundational LLM capabilities toward more sophisticated reasoning systems, with particular attention to evaluating and improving logical consistency, developing meta-reasoning benchmarks, and creating frameworks for automated alignment evaluation.

Dr. Guo has served as an Area Chair for NeurIPS 2025, reflecting his standing in the research community. His work on MR-Ben introduced a novel benchmark for evaluating System-2 thinking in LLMs, while his contributions to AVeriTeC created an important dataset for real-world claim verification with web evidence.

Dr. Guo actively seeks strong and motivated students to join his research group. His advising philosophy emphasizes fundamental research questions about LLM capabilities while maintaining practical relevance. Current projects focus on understanding the transition from System 1 to System 2 reasoning in language models, with significant implications for creating more reliable and trustworthy AI systems.

Dr. Guo leads research efforts focused on the intersection of knowledge representation and reasoning in large language models. His group develops new methodologies for evaluating and improving logical consistency, knowledge composition, and reasoning processes in AI systems. Current projects include investigating the transition from intuitive to deliberate reasoning in LLMs, developing efficient fine-tuning techniques like HydraLoRA, and creating frameworks for automated alignment evaluation in complex reasoning tasks.