Hello, I'm Junhao Cai, a second-year Ph.D. student in Computer Science and Engineering at Korea University in Seoul, advised by Professor Changhee Joo. I work on machine unlearning (making a trained model forget specific data on demand) and AI applications. Whether it's research or a project, I tend to get an end-to-end minimum viable product (MVP) running first, then improve it step by step.

So far that adds up to 6 papers, 3 of them at top-tier conferences such as NeurIPS and AAAI (with an oral at AAAI), and 8 patents. I'm also part of 3 national research projects funded by NRF and IITP.

My GPA is 100/100 in the Ph.D. and 97.1/100 in the M.S. I graduated from Yanbian University as an Engineering Innovation Class student with a B.E. in Computer Science and Technology (90.29/100), in the top 3 of my major. My honors include the Outstanding Paper Award from Korea University, the Samsung Electronic Scholarship (twice, top 1%), the National Inspirational Scholarship (a national-level award in China, top 1%) and the Excellent Student Scholarship in all eight undergraduate semesters.

Outside of research and work, I enjoy playing the piano and badminton.

I can communicate at a native level in both Chinese and Korean (TOPIK Level 6 in Korean) and I also work in English, so feel free to reach out in whichever language suits you, at or on LinkedIn.

안녕하세요, 고려대학교 컴퓨터학과 박사과정 2년차 채준호입니다. 주창희 교수님의 지도 아래 머신 언러닝(이미 학습된 AI 모델이 특정 데이터를 잊도록 만드는 기술)과 AI 응용을 연구하고 있습니다. 연구든 프로젝트든, 일단 처음부터 끝까지 한 번 돌아가는 최소 기능 제품(MVP)을 먼저 만들어 놓고 하나씩 다듬어 가는 편입니다.

지금까지 논문 6편을 냈고, 그중 3편은 NeurIPS, AAAI 같은 최상위 학회에 채택되었으며 AAAI에서는 Oral 발표로 선정되기도 했습니다. 특허도 8건 냈고, 한국연구재단(NRF)과 정보통신기획평가원(IITP)의 국가 연구과제 3건에도 참여하고 있습니다.

학점은 박사과정 100/100, 석사과정 97.1/100입니다. 학부는 중국 Yanbian University의 공학 혁신반(Engineering Innovation Class)에서 컴퓨터 과학 및 기술을 공부했고, 학점 90.29/100으로 전공 3등 안에 들었습니다. 고려대학교에서는 Outstanding Paper Award를 받았고, 학부 시절에는 Samsung Electronic Scholarship을 2회(상위 1%) 수상했고, 중국 국가급 장학금인 National Inspirational Scholarship(상위 1%)도 받았으며, Excellent Student Scholarship은 8학기 내내 빠짐없이 받았습니다.

연구실 밖에서는 피아노를 치거나 배드민턴을 치면서 머리를 식힙니다.

중국어와 한국어를 모두 원어민 수준으로 구사하고(한국어는 TOPIK 6급) 영어로도 일할 수 있으니, 편하신 언어로 이메일()이나 LinkedIn을 통해 연락 주세요.

您好,我是蔡俊豪,高丽大学(QS 排名第 52 位)计算机科学与工程系博士二年级在读,师从 Changhee Joo 教授(CCF A 类会议 INFOCOM 最佳论文奖获得者),主要研究机器遗忘(让训练好的 AI 模型按需忘掉指定的数据)和 AI 应用。不管是做研究还是做项目,我都倾向于先做出一个端到端跑通的最小可行产品(MVP),再一步步迭代优化。

到目前为止,我有论文 6 篇被录用,其中 3 篇被 NeurIPS、AAAI 等顶会录用,AAAI 的论文还入选了 Oral;另有专利 8 项,并正在参与 NRF、IITP 资助的韩国国家级科研项目 3 项。

成绩方面,我的博士 GPA 是 100/100,硕士是 97.1/100;本科就读于延边大学工科创新班(计算机科学与技术专业),GPA 90.29/100,专业排名前 3。荣誉方面,我博士期间获得了高丽大学的 Outstanding Paper Award,本科期间两次获得 Samsung Electronic Scholarship(前 1%),还获得了国家励志奖学金 National Inspirational Scholarship(前 1%),Excellent Student Scholarship 则 8 个学期一次都没落下。

科研和工作之余,我喜欢弹钢琴和打羽毛球。

中文和韩语我都能以母语水平交流(韩语有 TOPIK 6 级),英语也可以作为工作语言,无论您习惯哪种语言,都欢迎通过邮箱 或 LinkedIn 联系我。

💬 Thinking about joining our lab?

I am very fortunate to be advised by Prof. Changhee Joo of RAIN Lab.

If you’re reading this, you’re probably wondering what it’s like in our lab. Honestly? Our professor is amazing: not just brilliant at research, but also someone who really cares about his students. I never feel like I’m just another grad student to him, and he’s the kind of person who makes you want to push yourself while also making sure you don’t burn out. I’ve been here for a while now, and I can say this place has really shaped who I am as a researcher. It’s one of those environments where you actually look forward to coming in each day, so if you’re thinking about joining, I’d say go for it. You won’t regret it. =) 🍀

📖 Publications

  • NeurIPS 2026
    Retain-Neutral Surrogates for Min-Max Unlearning J Cai, D Kim, D Kim, SI Choi, C Jin, J Park, C Joo* CCF ABK+ IF 4Machine Unlearning
  • NeurIPS 2026
    Spectral Unlearning: Transformer Structure-Preserving Updates for Language Model SI Choi, J Cai, D Kim, C Joo* CCF ABK+ IF 4Machine Unlearning
  • IEEE COMMUN MAG 2026
    RXC: Runtime Extra Compression for Time-Critical Content Delivery DOI J Bae, H Kang, T Kim, T An, S Lee, J Cai, C Joo, K Lee SCIE, IF 8.2AI application
  • AAAI 2026 Oral
    HCF: Hierarchical Cascade Framework for Distributed Multi-Stage Image Compression DOI J Cai, T An, C Jin, SI Choi, J Park, C Joo CCF ABK+ IF 4AI application
  • COMPUT NETW 2025
    Economic and Strategic Perspectives on CDN Pricing: A Comprehensive Review DOI C Jin, J Cai, J Park, C Joo SCIECCF B
  • ICTC 2024 Oral
    Toward Scalable and Efficient Visual Data Transmission in 6G Networks DOI J Cai, T An, C Joo CCF none
  • Under Review ×6
    Conference papers, 2026

💡 Patents and Software Copyrights

* denotes equal contribution.

  • Patent 2026
    Device and Method for Retain-Orthogonal Updates Based Machine Unlearning for Continual Diffusion Concept Erasure PDF J Cai, C Joo
  • Patent 2025
    Device and Method for Model Performance-preserving Machine Unlearning for Lossy Training Data PDF C Joo, J Park*, J Cai*
  • Patent 2025
    Device and Method for Machine Unlearning Based on Per-sample Hessian Information PDF C Joo, SI Choi*, J Park*, J Cai*
  • Patent 2025
    Device and Method for Certified Machine Unlearning Using Dual Noise Injection PDF C Joo, SI Choi*, J Park*, J Cai*
  • Patent 2025
    Device and Method for Runtime Extra Compression for Time-Critical Content Delivery PDF C Joo, J Cai, T An
  • Patent 2025
    Device and Method of Graph-based Smart Offloading for Edge Computing PDF R Paul, J Cai, C Joo
  • Patent 2025
    Device and Method for Adaptive Image Compression for Distributed Environments Using Hierarchical Cascade Networks PDF J Cai, T An, C Joo AI application
  • Patent 2025
    Method for Differentiated Content Distributed Network Services on Market Structure PDF C Joo, S Lee, J Cai
  • Under Review ×4
    Patents, 2026
  • Software 2024
    Adaptive Image Compression System for Distributed Environments Using Hierarchical Cascade Networks PDF J Cai AI application

📝 Research Projects

Funded projects I take part in at RAIN Lab.

  • NRF
    Development of Core Technologies for Multimodal Machine Unlearning 2026 - 2031 멀티모달 머신 언러닝 핵심기술 개발
  • IITP
    High-level application-network integration programmable interface technology 2024 - 2028, Research Coordinator since 2024.07 고수준 응용-네트워크 연동 프로그래머블 인터페이스 기술 6G Industrial Technology DevelopmentNetwork Economics
  • NRF (ERC)
    Cognitive Augmentation Metacommunication Research Center 2022 - 2028, since 2023.09 인지증강 메타통신 연구센터AI ApplicationsNetwork for AI services (AI computing over network)

🏆 Honors and Awards

  • Ph.D.
    Global Leader Scholarship ×3 Top 1%Outstanding Paper Award
  • M.S.
    Global Leader Scholarship Top 1%Natural Science and Engineering Scholarship Top 5%Research Scholarship ×2Teaching Assistant Scholarship
  • B.E.
    Samsung Electronic Scholarship ×2 Top 1%National Inspirational Scholarship Top 1%Excellent Student Scholarship ×8 Top 1-5%

📑 Education

  • Ph.D. in Computer Science and Engineering, Korea University QS Ranking: 52nd, 2025.09 - Present GPA: 100/100.
  • M.S. in Computer Science and Engineering, Korea University QS Ranking: 52nd, 2023.09 - 2025.08 GPA: 97.1/100.
  • B.E. in Computer Science and Technology, Yanbian University 211 Project, 2019.09 - 2023.06 GPA: 90.29/100. Major Ranking: Top 3.

💼 Experience

  • Research Assistant (RA) and Teaching Assistant (TA), Korea University, 2025.09 - Present
  • Research Assistant (RA) and Teaching Assistant (TA), Korea University, 2023.09 - 2025.08
  • Major Representative and Undergraduate Research Assistant, Yanbian University, 2020.09 - 2023.06

🎓 Academic Activities

🌍 Languages

  • Chinese
    Native or bilingual proficiency
  • Korean
    Native or bilingual proficiency TOPIK Level 6 (The Highest Proficiency Level), 100% scores in listening and reading
  • English
    Professional working proficiency

🛠️ Skills

  • Coding
    Python, PyTorch, scikit-learn, CompressAI, etc.
  • Piano
    First Prize, National Piano Competition (China)National-Level Special Talent Student (China)First Prize, Jilin Province Piano Competition (China)
  • Badminton
    Special Talent Student (China)

Where visitors come from

Save a copy of this page