연구업적 및 과제성과

Under Revision and Under Review

YoungWook Kwon*, WanSoo Kim*, SuMin Oh, and HyunJin Kim, "DoTCoM: High-Performance Mobile Vision Transformers with Co-Bias," Multimedia Tools and Applications (Under Review)

SuMin Oh, WanSoo Kim, HyunJin Kim, "A Temporally Correlated Latent Exploration for Reinforcement Learning," Machine Learning (Under Review)

Min Kee Chang, DaeRyoung Shin, HyunJin Kim, Alberto A. Del Barrio, "Posit adders and multipliers with variable input and output exponent sizes." 2025 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS) (Under Review)

DaeRyoung Shin, Min Kee Chang, HyunJin Kim, Alberto A. Del Barrio, "AxAdderNet: Combining Approximate Computing to Quantized AdderNet" 2025 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS) (Under Review)

WanSoo Kim, YoungWook Kwon, HyunJin Kim, "Axial-Factorized Attention: A Fast and Efficient Approach to Real-Time Semantic Segmentation for MCU Deployment," IEEE Transactions on Intelligent Transportation Systems (Under Revision)

SuMin Oh, WanSoo Kim, HyunJin Kim, "A Novel Adaptive Testing Method using Convolutional Neural Networks," Journal of Semiconductor Technology and Science, (Under Review)

International Journal 

• SuMin Oh and HyunJin Kim, "Sample-Efficient Reinforcement Learning for Analog Circuit Optimization with Intrinsic Reward," Journal of Semiconductor Technology and Science, Accepted for Publication. (SCIE)  

International Conference

YoungWook Kwon*, WanSoo Kim*, HyunJin Kim, "HARD: Hardware-Aware lightweight Real-time semantic segmentation model Deployable from Edge to GPU," Asian Conference on Computer Vision 2024, https://doi.org/10.1007/978-981-96-0963-5_15, Lecture Notes in Computer Science, vol. 15478. Springer Singapore, pp. 3552~3569, Dec 8, 2024.

Sumin Oh, HyunJin Kim, "Late Breaking Results: Self-Motivated Agents for Analog Circuit Optimization via Intrinsic Reward," DAC 2025 Work-in-Progress. 

Sumin Oh, HyunJin Kim, "Late Breaking Results: POG: Parameter Optimization using Graph Neural Networks on Reinforcement Learning," DAC 2025 Work-in-Progress. 

Domestic Conference

Domestic Patent