1) Pruning & Compression for LLM
2) Pruning & Processing-In-Memory
2026-1) ToaSt: Token Channel Selection and Structured Pruning for Efficient ViT [ICML, BK21 IF=4]
Hyunchan Moon, Cheonjun Park* , and Steven L. Waslander
The International Conference on Machine Learning (ICML), Seoul, Korea, 2026
Acceptance rate: 26.6% , * : Corresponding author
2025-1) WINS: Winograd Structured Pruning for Fast Winograd Convolution [ICCV, BK21 IF=4, Highlight]
Cheonjun Park, Hyun Jae Oh, Mincheol Park, Hyunchan Moon, Minsik Kim, Suhyun Kim, Myung Kuk Yoon, and Won Woo Ro
The IEEE/CVF International Conference on Computer Vision (ICCV), Honolulu, Hawaii, USA, Oct. 19 - 23, 2025
Oral & Highlight acceptance rate: 2.9% [327/11152]
2025-2) Adversarial Purification via Super-Resolution and Diffusion [ICCV, BK21 IF=4]
Mincheol Park, Cheonjun Park, Seungseop Lim, Mijin Koo, Hyunwuk Lee, Won Woo Ro, and Suhyun Kim
The IEEE/CVF International Conference on Computer Vision (ICCV), Honolulu, Hawaii, USA, Oct. 19 - 23, 2025
Acceptance rate: 24% [2701/11152]
2024-1) DEPrune: Depth-wise Separable Convolution Pruning for Maximizing GPU Parallelism [NeurIPS, BK21 IF=4]
Cheonjun Park, Mincheol Park, Hyunchan Moon, Myung Kuk Yoon, Seokjin Go, Suhyun Kim, and Won Woo Ro
The 38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024), Vancouver, Canada, Dec. 9 - 15, 2024
Acceptance rate: 25.75% [4035/15671]
2024-2) Reprune: Channel pruning via kernel representative selection [AAAI, BK21 IF=4]
Mincheol Park, Dongjin Kim, Cheonjun Park, Yuna Park, Gyeong Eun Gong, Won Woo Ro, and Suhyun Kim
The 37th Association for the Advancement of Artificial Intelligence (AAAI-24), Vancouver, Canada, Mar. 24 , 2024
Acceptance rate: 23.75% [2342/12100]
2023-1) Balanced column-wise block pruning for maximizing GPU parallelism [AAAI, BK21 IF=4, Oral]
Cheonjun Park, Mincheol Park, Hyun Jae Oh, Minkyu Kim, Myung Kuk Yoon, Suhyun Kim, and Won Woo Ro
The 37th Association for the Advancement of Artificial Intelligence (AAAI-23), Washington DC, USA, Feb. 07 - 14, 2023
Oral acceptance rate: 10.8% [952/8777]
2019-1) Efficient dilated-winograd convolutional neural networks [ICIP]
Minsik Kim, Cheonjun Park, Sungjun Kim, Taeyoung Hong, and Won Woo Ro
2019 IEEE International Conference on Image Processing (ICIP), 2711-2715
soon ...