Hyundong Jin, Ph.D. (He/Him)
wlsgusehd@gmail.com
Research expertise in continual learning, enabling scalable knowledge retention and adaptive capabilities by mitigating task interference across diverse domains and modalities. Explore research directions including multimodal learning, resource-efficient learning, and multimodal large language models, ultimately aiming toward Artificial General Intelligence (AGI).
News!
Overall Winner 🏆, CVPR 2026 Foundational FSOD Challenge
Won the Foundational Few-Shot Object Detection challenge using Superb AI's model ZERO.
Joined Superb AI as a Machine Learning Engineer 💼
Serving as an Expert Research Personnel (alternative military service).
Received my Ph.D. in Computer Science and Engineering 🎓
Chung-Ang University
One paper accepted to CVPR 2026 🎉
Continual unlearning for large vision-language models.
One paper accepted to ICLR 2026 🎉
XIL: Cross-Expanding Incremental Learning.
One paper accepted to Neural Networks 🎉
Exploration and Exploitation in Continual Learning.
One paper accepted to ICCV 2025 🎉
Instruction-grounded visual projectors for continual learning of generative VLMs.
One paper accepted to ICCV 2023 🎉
Growing a brain with sparsity-inducing generation for continual learning.
One paper accepted to ECCV 2022 🎉
Helpful or harmful: inter-task association in continual learning.
Publications
Instruction-Grounded Visual Projectors for Continual Learning of Generative Vision-Language Models
IEEE International Conference on Computer Vision (ICCV), 2025
Also presented at Korea Robotics Society (KROS) Physical AI Workshop, 2026
Gating Mechanism in Deep Neural Networks for Resource-Efficient Continual Learning
IEEE Access, 2022
Ongoing Works
Mind the Interference: Towards Robust Continual Learning Across Modalities
To be updated
Action-incremental Learning for Temporal Action Segmentation
To be updated
Project Experiences
Multi-Modal Continual Learning with Context Understanding
Funded by National Research Foundation
Time-Series Action Prediction and Segmentation
Funded by HD Hyundai Construction Equipment
Learning Transferable Task Knowledge and Planner for Service Robots
Funded by Samsung Research Funding & Incubation Center
Development of AI for Self-Improving Competency-Aware Learning
Funded by IITP
Automated Deep Learning Technology for Multi-Task Learning
Funded by National Research Foundation
Others
Awards & Honors
- Overall Winner, CVPR 2026 Foundational FSOD Challenge · The 6th Workshop of Open World Vision, 2026
- Grand Prize, Big Data Utilization Contest · Doosan Enerbility, 2023
- Excellence Prize, Big Data Utilization Contest · HD Hyundai XiteSolution, 2023
Patents
- Apparatus and Method for Continuous Learning of Neural Networks· Republic of Korea. 10-2023-0156623
- A Neural Network Apparatus and Neural Network Learning Method for Performing Continuous Learning Using a Correlation Analysis Algorithm Between Tasks· Republic of Korea. 10-2022-0101187
Invited Talks
- AhnLab · Continual Learning, 2023
- Korean Computer Vision Society (KCVS) · Continual Learning session, 2023
- Korean Artificial Intelligence Association (KAIA) & NAVER · CV / NLP session, 2022
Teaching Experiences
- Machine Learning, Teaching Assistant (TA) · Chung-Ang University, 2024
- Advanced Artificial Intelligence, Teaching Assistant (TA) · Chung-Ang University, 2023
- Capstone Design, Teaching Assistant (TA) · Chung-Ang University, 2021
- Visual Intelligence and its Applications · ETRI, 2020
- Algorithms, Teaching Assistant (TA) · Chung-Ang University, 2020
Academic Services
- Conference Reviewer · CVPR, ICCV, ECCV, AAAI, ICLR, WACV, BMVC
- Journal Reviewer · TNNLS, Neural Networks, Neurocomputing