I am Dayoung, a Ph.D. student at the Graduate School of Soongsil University, conducting my research under the supervision of Professor Seong Heum Kim in the Visual Intelligence and Platform (VIP) Lab. My research primarily focuses on hardware-efficient artificial intelligence, with particular emphasis on Efficient Deep Learning, Pruning, Large Vision-Language Models (LVLMs), and Multimodal Large Language Models (MM-LLMs). Recently, I have been actively exploring LVLMs and advanced token pruning techniques to enhance the efficiency and performance of large-scale multimodal systems.

Research Interests

• Computer Vision
• LVLM (Large Vision-Language Models)
• MM-LLM (Multimodal Large Language Models)
• Efficient Deep Learning
• Model Pruning

Publications

One-Cycle Structured Pruning via Stability-Driven Subnetwork Search

Ghimire, D., Kil, D., Jeong, S., Park, J., & Kim, S. H. (2025). One-cycle Structured Pruning with Stability Driven Structure Search. arXiv preprint arXiv:2501.13439.

A Study of Structure Pruning for Hybrid Neural Networks

D. Ghimire, D. Kil and S. -h. Kim, "A Study of Structured Pruning for Hybrid Neural Networks," 2024 24th International Conference on Control, Automation and Systems (ICCAS), Jeju, Korea, Republic of, 2024, pp. 1110-1113, doi: 10.23919/ICCAS63016.2024.10773379.

Lightweight Room Layout Estimation using a Single Panoramic Image

D. Kil and S. -H. Kim, "Lightweight Room Layout Estimation using a Single Panoramic Image," 2022 22nd International Conference on Control, Automation and Systems (ICCAS), Jeju, Korea, Republic of, 2022, pp. 1951-1953, doi: 10.23919/ICCAS55662.2022.10003901.

단일 파노라마 입력의 실내 공간 레이아웃 복원 모델 경량화 [KCI]

Dayoung Kil and Seong-heum Kim. (2022). Lightweight Deep Learning for Room Layout Estimation with a Single Panoramic Image. Journal of Institute of Control, Robotics and Systems, 28(10), 868-873.

A Survey on Efficient Convolutional Neural Networks and Hardware Acceleration

Ghimire, D.; Kil, D.; Kim, S.-h. A Survey on Efficient Convolutional Neural Networks and Hardware Acceleration. Electronics 2022, 11, 945. https://doi.org/10.3390/electronics11060945

Awards

4th, Track1 - Image classification for different lighting conditions and styles

최우수상(ITTP 원장상), 차량용 객체복합 상태 인식

우수논문상, 단일 파노라마 입력의 실내 공간 레이아웃 복원 모델 경량화

장려상, Face Out : 얼굴 인식을 통한 사람 얼굴 모자이크 처리 웹

장려상, Face Out : 얼굴 인식을 통한 사람 얼굴 모자이크 처리 웹

우수상

Patents

• 길다영, 김성흠, “단일 파노라마 입력의 실내 공간 레이아웃 복원 모델 경량화,” C-2022-036602 (소프트웨어등록)

Educations

Mar 2024 - Feb 2027

Doctor of Intelligent Semiconductors Soongil University, Seoul, South Korea

Mar 2022 - Feb 2024

Master of Intelligent Systems Soongil University, Seoul, South Korea

Mar 2018 - Feb 2022

Bachelor of AI Convergence Soongil University, Seoul, South Korea

Mar 2014 - Feb 2017

Kojan High School (General Program) Ansan, Gyeonggi-do, South Korea