About
Yicheng Sun is a PhD Student in the ASI Lab. of Department of Computer Science, City University of Hong Kong.
Yicheng Sun researches LLM-driven software engineering, focusing on automated program repair and software log analysis. In RECAP, he shows that successful LLM-repair patches are systematically more verbose than developer patches, the median approach producing 121.78% more total changes, and proposes post generation refinement that cuts average total changes from +242.14% to +4.24% while preserving or improving resolution. He co-authors the first empirical study of Model Context Protocol adoption, interviewing 20 practitioners across eight companies. His R2ComSync improves code-comment synchronization via in-context learning with ensemble hybrid retrieval and multi-turn re-ranking, outperforming state-of-the-art across five LLMs and three Java/Python datasets.
Research Interests
Publications (29)
-
Technology2026ConferenceTechnology in Education (Springer LNCS)
-
APSEC2025
-
APSEC2025
-
ASE2025
-
COMPSAC2024