News
- Prof. Han-Lim Choi’s Research Team, Receives Best Paper Award at the 2025 Journal of KIMST
- 관리자 |
- 2026-06-17 14:15:03|
- 29
- 2026-06-17 14:15:03|
On June 10, 2026, a research team from our department's Laboratory for Information and Control Systems (LiCS), consisting of Ph.D. candidate Su-Jeong Park, M.S. alumnus Daegyeong Roh, and Ph.D. candidate Minchae Kim, received the Best Paper Award at the 2025 Journal of KIMST for work conducted under the industry-academia collaboration project "Research on AI-Based Threat Target Classification Algorithms" with Hanwha Systems Co., Ltd.
The awarded paper, "Out-of-Distribution-Aware Contrastive Learning for Trajectory-Based Threat Identification," proposed an AI model for real-time threat classification using radar trajectory data. Conventional classifiers have a key limitation: even when a target does not belong to any predefined threat type, they tend to force it into one of the known types. By applying supervised contrastive learning, the team enabled the model to flag such unknown targets as "unclassified," offering a way to reduce misidentification and unnecessary resource expenditure in real-world battlefield environments.
The awarded paper, "Out-of-Distribution-Aware Contrastive Learning for Trajectory-Based Threat Identification," proposed an AI model for real-time threat classification using radar trajectory data. Conventional classifiers have a key limitation: even when a target does not belong to any predefined threat type, they tend to force it into one of the known types. By applying supervised contrastive learning, the team enabled the model to flag such unknown targets as "unclassified," offering a way to reduce misidentification and unnecessary resource expenditure in real-world battlefield environments.
| Attach File |
|---|