Zhiyu Fan

Bio

Zhiyu Fan is a final year PhD candidate in the School of Computing at the National University of Singapore, advised by Prof. Abhik Roychoudhury. He obtained my Bachelor's degree in Computer Science from the Southern University of Science and Technology in China, where he was advised by Prof. Shin Hwei Tan.

He is broadly interest in Software Engineering and Generative AI, but his research has been focused on the intersections of automated program repair, automatic programming, and computer science education. His research work on program repair for intelligent tutoring (TOSEM22, ISSTA23), program repair of AI programs (ICSE23), explorations on AI (i.e. LLM) programming (ISSTA24a, ISSTA24b) have been published in top-tier software engineering venues. His has led the development of Intelligent Tutoring System to boost Computer Science education at scale via SE and LLM techniques, which has been deployed in NUS. Detail can be found at [Projects].

Education

• National University University, Singapore 2019 - Present
  Ph.D student (School of Computing)
  Advisor: Prof. Abhik Roychoudhury
• Southern University of Science and Technology, China 2015 - 2019
  B.E., (Dept of Computer Science)
  Advisor: Prof. Shin Hwei Tan

Publications

• [ISSTA 2024] AutoCodeRover: Autonomous Program Improvement Paper

  Yuntong Zhang, Haifeng Ruan, Zhiyu Fan, Abhik Roychoudhury

  International Symposium on Software Testing and Analysis (ISSTA), 2024

  (One line abstract) Natural language as program specification, equip LLMs with debugging capability to resolve real-world Github Issues

• [ISSTA 2024] Oracle-guided Program Selection from Large Language Models Paper

  Zhiyu Fan, Haifeng Ruan, Sergey Mechtaev, Abhik Roychoudhury

  International Symposium on Software Testing and Analysis (ISSTA), 2024

  (One line abstract) LLMs can effectively work as oracle to select correct AI-programs regard natural language description

• [arXiv] Intelligent Tutoring System: Experience of Linking Software Engineering and Programming Teaching Paper

  Zhiyu Fan, Yannic Noller, Ashish Dandekar, Abhik Roychoudhury

  arXiv preprint, 2023

  (One line abstract) An modular and extensible feedback generation system that consist of advanced research ideas from APR and LLM

• [ISSTA 2023] Concept-based Automated Grading of CS-1 Programming Assignments Paper

  Zhiyu Fan, Shin Hwei Tan, Abhik Roychoudhury

  International Symposium on Software Testing and Analysis (ISSTA), 2023.

  (One line abstract) We proposed the notion of a concept graph for grading and providing feedback for programming assignments

• [ICSE 2023] Automated Repair of Programs from Large Language Models Paper

  Zhiyu Fan, Xiang Gao, Martin Mirchev, Abhik Roychoudhury, Shin Hwei Tan

  45th IEEE/ACM International Conference on Software Engineering (ICSE), 2023.

  (One line abstract) The first systematic study of LLM for APR and APR of LLM-generated programs

• [TOSEM 2022] Verifix: Verified Repair of Programming Assignments Paper

  Umair Z. Ahmed*, Zhiyu Fan*, Jooyong Yi, Omar I. Al-Bataineh, Abhik Roychoudhury.

  ACM Transactions on Software Engineering and Methodology (TOSEM), 2022.

  (One line abstract) Verified repair patches (guaranteed correctness and minimal changes) as feedback to student, avoiding misleading guidance.

Services

• Reviewer for TOSEM 2024
• PC Member for APR Workshop @ ICSE 2024
• Reviewer for SIGCSE 2024
• Reviewer for TSE 2023

Selected Awards

• NUS Research Achievement Award 2023
• NUS Research Scholarship 2019 - 2024
• SUSTech Outstanding Graduate Student Award 2019
• ACM SIGSOFT CAPS Grant 2019

Teaching Experience

• Spring 2024 CS2040S - Data Structures and Algorithms Teaching Assistant
• Spring 2023 CS3213 - Foundations of Software Engineering Head Teaching Assistant
• Spring 2022 CS3213 - Foundations of Software Engineering Hdea Teaching Assistant
• Fall 2020 CS4211 - Formal Method for Software Engineering Teaching Assistant