AI-Powered Software Engineering (SELab)

@ HanKyong National University (HKNU)

Our News

[Now Recruiting] Call for Postdoctoral and Graduate Researchers

- SELAB@HKNU 연구실에서는 AI와 소프트웨어 공학 분야에서 탁월한 역량을 갖춘 박사 후 연구원(Post-Doctoral) 및 대학원생/학-석사연계생을 모집합니다. 저희 연구팀은 AI 기술을 통해 소프트웨어 개발의 패러다임을 혁신하고, 실질적인 영향을 미칠 수 있는 고도화된 기술을 개발하는 데 주력하고 있습니다.
- RAG-sLLM, LLMs, SLM, LMM 기술뿐만 아니라 GPU 최적화와 GPU 양자화도 즐기면서 경량화된 대규모 언어 모델을 탐구하고 즐기실 분들을 환영합니다.
- 또한, 성실함과 협업 능력, 그리고 높은 수준의 전문성을 갖춘 인재를 찾고 있습니다. 다양한 연구 기회를 탐색하고 더 자세한 정보를 원하시면, 양근석 교수님에게 문의하시거나 연구실 웹사이트를 방문해 주세요: [View Details]

Join us at the 40th ACM Symposium on Applied Computing (SAC) in Sicily, Italy, from March 31 to April 4, 2025!

- Discover groundbreaking ideas and cutting-edge technologies at ACM SAC 2025. Connect with leading researchers and practitioners from around the world, attend insightful presentations, participate in interactive workshops, and network with peers in the beautiful setting of Sicily. This is a unique opportunity to shape the future of applied computing. Don't miss out on the chance to be part of this prestigious event. [View Details]

[Notice] 사용자 중심 소프트웨어 서비스 개발 프로젝트 협력 환영

- SELAB@HKNU에서는 사용자 중심의 소프트웨어 서비스 연구 개발 프로젝트에 함께할 파트너를 기다리고 있습니다. 저희 연구실은 SW 개발/테스트 전문성과 함께 AI, 검색 증강 경량 거대 언어 모델 (RAG-sLLM), 거대 언어 모델 (LLMs), 소형 언어 모델 (SLM), 대형 멀티모달 모델 (LMMs), 빅데이터, 클라우드 기술, 정보시스템 감리 등 다양한 분야에서 뛰어난 전문 지식을 보유하고 있습니다. 특히, GS 인증, 감사 준비가 완료된 결과물, AI 서비스, 데이터 과학 분야에서 협력할 파트너를 찾고 있습니다. 저희와 함께 혁신적인 프로젝트(NIA/NIPA/IITP/중기청/K-Startup 등)를 성공적으로 이끌어갈 파트너를 찾고 있으니, 많은 관심과 문의 부탁드립니다.

[News] Our Recent Achievements
[2024-12-20] 우리 연구실에서 2024 한국소프트웨어종합학술대회의 인공지능-소프트웨어공학 트랙에 포스터(텍스트 유사도 기반 RAG 와 sLLM을 활용한 소프트웨어 보안 버그 리포트 템플릿 생성 및 예측 방법, 여수EXPO)을 발표합니다.
[2024-12-20] 우리 연구실에서 2024 한국소프트웨어종합학술대회의 소프트웨어공학 트랙에 정규 논문(The Effect of Feature Selection on the Performance of Duplicate Bug Report Detection A Scenario-Based Analysis, 여수EXPO)을 발표합니다.
[2024-06-28] 우리 연구실에서 2024 한국컴퓨터종합학술대회의 소프트웨어공학 트랙에 정규 논문(Enhancing Security Bug Report Prediction with Severity-Based Feature Selection through LSTM Algorithm / 제주 ICC) 발표를 하였습니다.
[2024-04-08] 우리 연구실에서 The 39th ACM/SIGAPP Symposium On Applied Computing (Avila, Spain) 국제학술대회에 참석하였습니다.
[2023-12-22] 우리 연구실에서 2023년 한국소프트웨어종합학술대회(부산 벡스코)에서 "소프트웨어 진화에 거대 언어 모델 적용"을 발표하였습니다.

Advancing RAG+sLLM and RAG+LMM Research

AI-Driven Code Synthesis

SELAB@HKNU is at the forefront of cutting-edge RAG+sLLM-powered development tools that go beyond traditional code completion. Our specialized models, meticulously trained on vast and diverse codebases, autonomously generate accurate, context-aware code snippets that adhere to rigorous industry standards and best practices. By automating complex and repetitive coding tasks, our solutions enhance development efficiency, reduce error rates, and empower developers to focus on high-impact, strategic problem-solving.

Code That Speaks Your Language

SELAB@HKNU is leading a transformative shift in software development by enabling seamless natural language interaction with code. Our advanced RAG+sLLM-based interfaces empower developers to express programming intentions in plain English, bridging the gap between human thought and machine logic. This pioneering technology streamlines workflows for seasoned developers and democratizes coding by lowering entry barriers for aspiring programmers, fostering a more inclusive and innovative development ecosystem.

Self-Documenting Code

SELAB@HKNU is transforming code maintainability and knowledge sharing with advanced RAG+sLLM-driven solutions. Our models autonomously generate comprehensive, up-to-date documentation and insightful comments directly from the codebase, eliminating the manual, error-prone documentation process. This ensures code remains thoroughly explained, accessible to team members, and optimally prepared for efficient onboarding, greatly enhancing long-term project sustainability and collaborative productivity.

Seeing Code in a New Light

SELAB@HKNU is pioneering the future of code understanding through cutting-edge research in Large Multimodal Models (LMMs). Our approach reimagines code as more than just text, transforming it into interactive visualizations, structural diagrams, and other intuitive representations. This multimodal perspective enables developers to gain deep insights into complex code architectures, quickly identify potential issues, and collaborate with unmatched clarity. By bridging the gap between machine code and human comprehension, we empower developers to create higher-quality software solutions.

Visualizing Code Creation

SELAB@HKNU is redefining code generation by transforming it into an interactive visual experience. Our advanced LMM-powered tools go beyond producing functional code, generating accompanying diagrams, flowcharts, and visual elements that reveal the code's underlying logic. This visual-focused approach significantly enhances code comprehension, streamlining review, debugging, and maintenance processes to ensure higher software quality and maintainability.

AI-Powered Debugging

SELAB@HKNU is advancing debugging technology with state-of-the-art LMM-powered tools that enhance accuracy and efficiency in error resolution. Our innovative approach analyzes code alongside multimodal artifacts such as logs and execution traces, enabling precise error detection and intelligent fix recommendations. This not only speeds up the debugging process but also empowers developers to proactively prevent future issues, fostering the creation of more resilient and reliable software solutions.

Cultivating AI Talent in RAG+sLLM and RAG+LMM

Small Large Language Models (sLLM)

SELAB@HKNU is pioneering advanced methodologies for training and fine-tuning sLLMs on vast, high-quality code datasets. Our models excel at generating accurate code, completing code snippets, translating between programming languages, and understanding natural language instructions for a variety of software development tasks. This cutting-edge work expands the capabilities of sLLMs, setting new standards in AI-driven software engineering.

Large Multimodal Models (LMMs)

SELAB@HKNU is at the forefront of leveraging LMMs to represent and interpret code across multiple modalities, including text, images, and structural diagrams. Our research advances techniques for multimodal code generation, completion, and debugging, unlocking deeper insights by utilizing the rich contextual information inherent in diverse representations. This approach sets new benchmarks in code comprehension and functionality, driving innovation in AI-powered software engineering.

Retrieval-Augmented Generation (RAG)

SELAB@HKNU is pioneering sophisticated techniques in Retrieval-Augmented Generation (RAG) to dynamically integrate external knowledge into language models. Our research enhances these models' ability to generate accurate, contextually rich responses for complex tasks across diverse domains, including software engineering and multimodal content creation. This work drives forward the precision and adaptability of AI-powered solutions, setting new standards in responsive and context-aware technology.

Big Data Techniques for Code Analysis

SELAB@HKNU utilizes state-of-the-art big data techniques to process and analyze vast code repositories, uncovering critical insights that drive the advancement of RAG+sLLM and LMMs in software engineering. Our approach enables a deeper understanding of code patterns and optimizations, accelerating the development of intelligent, high-performance AI models tailored for complex software tasks.

Cloud-Native Techniques for Model Deployment

SELAB@HKNU leverages advanced cloud-native technologies to design robust, scalable infrastructure for training, deploying, and managing RAG+sLLM and RAG+LMM models. Our cloud-based solutions ensure efficient and reliable access for developers and researchers, streamlining model integration and accelerating innovation in AI-driven software development.

Open Source Contributions

SELAB@HKNU is committed to advancing open-source RAG+sLLM and RAG+LMM initiatives, fostering a collaborative ecosystem that drives innovation and knowledge sharing within the AI and software engineering communities. Our active contributions accelerate progress in state-of-the-art technologies, empowering developers and researchers globally.

Frequently Asked Questions

1. Benefits for Our Researchers

Our lab provides an exceptional environment for graduate students to excel in practical AI, RAG+sLLM, RAG+LMM, and advanced software development, preparing them as leaders in these fields. As a member of our team, you will:

- Develop Expertise in Applied AI: Gain hands-on experience and deep technical knowledge in using AI to automate software artifact generation through pioneering research and development projects.

- Engage in High-Impact Research: Contribute to advisor-led projects that push the boundaries of AI-driven applications in software engineering, aligned with your specific interests and expertise.

- Stay at the Forefront of Innovation: Attend national and international conferences and technical seminars, connecting with industry leaders and staying up-to-date on the latest advancements in AI and software engineering.

2. Supportive Lab Environment

Our lab cultivates a dynamic and collaborative research environment, emphasizing impactful outcomes and personalized growth for each researcher:

- Insightful Lab Meetings: Participate in engaging discussions on the latest advancements in AI, RAG+sLLM, RAG+LMMs, and their practical applications in software engineering, enhancing both knowledge and research acumen.

- Comprehensive Lab Seminars: Attend structured knowledge-sharing sessions to deepen your expertise in AI, RAG+sLLM, RAG+LMMs, and emerging trends in software engineering.

- Personalized Career Development Support: Benefit from one-on-one guidance tailored to your career goals, including preparation for coding assessments and interviews in AI and software engineering fields.

- Flexible, Results-Oriented Environment: Thrive in a flexible work culture that supports remote research, empowering you to focus on innovation and professional growth without traditional commuting constraints.

3. Our Vision

SELAB@HKNU is united by a vision to transform the software engineering landscape through groundbreaking AI-powered solutions:

- Pioneering Research Excellence: We are dedicated to conducting rigorous, pioneering research in applying AI, RAG+sLLM, and RAG+LMMs to software engineering, continually pushing the frontiers of knowledge and innovation.

- Cultivating World-Class Talent: Our mission is to develop the next generation of researchers and developers with the expertise, creativity, and vision to lead advancements in AI-driven software engineering.

- Driving Industry Transformation: We aspire to deliver high-impact solutions and foster thought leaders who will revolutionize the software engineering industry by harnessing AI, RAG+sLLM, and RAG+LMMs to enhance productivity, quality, and innovation.