Projects

Research projects on AI for software engineering and software maintenance and evolution.
Bug Report Quality Assessment and Enhancement

LLM-powered tools (AstroBR, BugScribe, and BURT++) that leverage app context, such as GUI screenshots and interaction traces, to automatically assess, enhance, and interactively guide the creation of mobile app bug reports.

[ICPC'25] [BugScribe] [BURT++]

Bug Report Information for AI Coding Agents

Studying how bug report information (e.g., reproduction steps, GUI information, and code localization) affects LLM coding agents in repairing UI-centric Android bugs.

[Preprint]

Issue Resolution in Practice

Characterizing real-world issue resolution workflows through qualitative analysis of issue reports and developer interviews, for both traditional and AI/ML software systems.

[ICSE'25] [ICSME'26] [ICSE'26]

Solution Identification in Issue Discussions

Comparing ML models and LLMs via embeddings, prompting, and fine-tuning to automatically identify solution-related content in developer discussions.

[Preprint]

Buggy UI Localization

Uni/multimodal deep learning models (e.g., SBERT, CLIP, BLIP) for identifying buggy UI screens and components from natural language bug descriptions.

[ISSTA'24]

Code Change Rationale

Multi-document extraction and generation of code change rationale from commits, pull requests, and issues.

[Preprint]

SPRINT: Issue Report Management

An LLM-powered GitHub plugin that finds similar issues, predicts issue severity, and suggests code files to modify.

[MSR'25] [Code]