Welcome to the 2025 Next Generation of CodeLLMs and Their Applications

Over the last decade, code generation techniques have rapidly grown at a tremendous rate, on the basis of large language models (LLMs). Today, we stand at the frontier of a new era where CodeLLMs are not just tools for code generation and completion, but intelligent collaborators capable of reasoning, debugging, and co-designing complex software systems.

This event brings together leading researchers and practitioners who are shaping the next generation of AI-powered Software Engineering. We invite distinguished speakers from around the world, each bringing innovative perspectives and groundbreaking ideas to advance the frontiers of CodeLLM research and intelligent software development.

We hope you enjoy the state-of-the-art research, inspiring new ideas, and valuable opportunities to connect with fellow researchers throughout this workshop !!!

Venue

Woonoh IT Hall @ Korea University

Kakao Map: https://place.map.kakao.com/1991433382

Google Maps: https://maps.app.goo.gl/whphTirpRkecZij96

Open Street Map: https://www.openstreetmap.org/#map=18/37.584790/127.028743

Transportations

Basically, you need to get to Anam Station (Metro Line 6).

• By public transportation from the conference hotel

• By taxi or Uber from the conference hotel

Direction

From Anam Station (Line 6)

(Kakao Map)

(Google Maps)

(Naver Map)

Program at a glance

Day 1, 16th November, 2025 (Sunday)

Session 1

Venue: Room 609, Woonoh IT Hall

• 14:00pm / Opening
• 14:30pm / Tobias Kiecker / Dependency Debloating via Method Replacement (with LLMs)
• 15:00pm / Honghao Tan / Coverage-Based Harmfulness Testing for LLM Code Transformation
• 15:30pm / Zhou Yang / Advancing Software, Society, and Security with Trusted Intelligence
• 16:00pm / Chaiyong Ragkhitwetsagul / Social Media Reactions to Open Source Promotions: AI-Powered GitHub Projects on Hacker News
• 16:30pm / Jieke Shi / Optimizing and Fortifying AI Software via Program Synthesis
• 17:00pm / Hengcheng Zhu / LSPFuzz: Hunting Bugs in Language Servers
• 17:30pm / Discussion & Closing

Day 2, 20th November, 2025 (Thursday)

Session 1

Venue: Room 520, Woonoh IT Hall

• 10:00 / Opening
• 10:30 / Zhongxin Liu / From LLM-Driven Fault Localization to Fault-Aware Code Generation
• 11:00 / Sergey Mechtaev / Automated Repair of Ambiguous Problem Descriptions for LLM-Based Code Generation
• 11:30 / Xiaodong Gu / On understanding complex software repository

Lunch

• 12:00 / Lunch

Session 2

Venue: Room 609, Woonoh IT Hall

• 13:30 / Weiyi Shang / Towards Pragmatic Pre-processing of Logs
• 14:00 / Zhenhao Li / Defects4Log: Benchmarking LLMs for Logging Code Defect Detection and Reasoning
• 14:30 / Zhensu Sun / AI-friendly code representation in an era of LLMs

Session 3

Venue: Room 616, Woonoh IT Hall

• 15:00 / Qiang Hu / Defects4C: Benchmarking Large Language Model Repair Capability with C/C++ Bugs
• 15:30 / Shou Liu / Can Mamba Be Better? An Experimental Evaluation of Mamba in Code Intelligence
• 16:00 / Tiezhu Sun / RAML: Toward Retrieval-Augmented Localization of Malicious Payloads in Android Apps
• 16:30 / Dianshu Liao / Navigating the Labyrinth: Path-Sensitive Unit Test Generation with Large Language Models
• 17:00 / Haochen Gong / Towards Context-aware Mobile Privacy Notice: Implementation of A Deployable Contextual Privacy Policies Generator
• 17:30 / Junda He / From Code to Courtroom: LLMs as the New Software Judges

Dinner

(Tentative)