AI Acceleration

Hardware and architecture for efficient deep learning

AI Acceleration

Overview

As AI models continue to grow in scale and complexity, achieving high performance and energy efficiency has become a critical challenge in modern computing systems. We investigate AI acceleration techniques across computing architectures and large-scale AI systems. Our research includes energy-efficient AI computing architectures, communication optimization, workload balancing, and topology-aware computation and data placement for efficient distributed AI processing.

Selected Publications

Top-tier PACT
ScaleMoE: A Fast and Scalable Distributed Training Framework for Large-Scale Mixture-of-Experts Models
Seo Hong choi, Huize Hong, Tae Hee Han and Joon Sung Kim
2025 34th International Conference on Parallel Architectures and Compilation Techniques (PACT), Nov. 03, 2025
#AI Accel #MoE #Distributed Training Abstract
ISOCC
Reducing Data Transfer Overhead with CXL-Based Near-Data Processing for LLM Inference
Woo Hyun Kim, Han Sol Kim, Tae Hee Han
2025 22nd International SoC Design Conference (ISOCC), Oct. 15, 2025
#Memory-Centric #AI Accel #NDP #CXL #LLM Abstract
J. Supercomput.
LNBN: layer-flexible non-blocking bypass network-on-chip for accelerating DNN inference
Suk Bong Kang, Won Hyeok Kim, Tae Hee Han
The Journal of Supercomputing, Volume 81, article number 1434, Oct. 10, 2025
#AI Accel #NoC #Bypass Network #DNN Inference Abstract
IEEE Access
BTCP:Binary Temporal Convolutional Network-based Data Prefetcher for Low Inference Latency and Storage Overhead
Chang Ho Ryu and Tae Hee Han
IEEE Access, Vol. 13, pp. 115048-115062, Jul. 02, 2025
#Memory System #AI Accel #Prefetching #ML-Based Abstract
PLOS ONE
Deep learning-based material decomposition of iodine and calcium in mobile photon counting detector CT
Kwanhee Han,Chang Ho Ryu,Chang-Lae Lee,Tae Hee Han
PLOS ONE, Jul. 26, 2024
#AI Accel #Medical Imaging Abstract
ICAIIC
RISC-V R-Extension: Advancing Efficiency with Rented-Pipeline for Edge DNN Processing
Won Hyeok Kim, Hyeong Jin Kim and Tae Hee Han
2024 International Conference on Artificial Intelligence in Information and Communication (ICAIIC), pp. 800-805, Osaka, Japan, Feb. 2024
#AI Accel #SoC Design #RISC-V #Edge DNN Abstract
Featured Article IEEE Access
RL-Based Cache Replacement: A Modern Interpretation of Belady’s Algorithm With Bypass Mechanism and Access Type Analysis
Ho Jung Yoo, Jeong Hun Kim and Tae Hee Han
IEEE Access, Vol. 11, pp. 145238-145253, Dec. 2023
#Memory System #AI Accel #Cache #ML-Based Abstract
IEEE Access
SFAO: Sign-Flipping-Aware Optimization for Early-Stopping of Binarized Neural Networks
Ju Yeon Kang, Chang Ho Ryu, Suk Bong Kang, and Tae Hee Han
IEEE Access, Vol. 11, pp. 128306-128315, Nov. 2023
#AI Accel #Quantization #BNN Abstract
IEEE Access
Binarized Neural Network with Parameterized Weight Clipping and Quantization Gap Minimization for Online Knowledge Distillation
Ju Yeon Kang, Chang Ho Ryu, and Tae Hee Han
IEEE Access, Vol. 11, pp. 8057-8064, Jan. 2023
#AI Accel #Quantization #Distillation Abstract
Electronics
CSGN: Combined Channel- and Spatial-wise Dynamic Gating Architecture for Convolutional Neural Networks
Sang Min Hyun, Chang Ho Ryu, Ju Yeon Kang, Hyun Jo Lim, and Tae Hee Han
Electronics, Vol. 11, No. 17, Aug. 2022
#AI Accel #CNN #Dynamic Gating Abstract
IEEE Access
On-the-fly Lowering Engine: Offloading Data Layout Conversion for Convolutional Neural Networks
Min Gu Kang, Sang Min Hyun, Tae Hee Han, Seok In Hong, and Jung Rae Kim
IEEE Access, Vol. 10, pp. 79370-79746, Jul. 2022
#Memory-Centric #AI Accel #CNN #Dataflow #Offloading Abstract
IEEE Access
Task Parallelism-Aware Deep Neural Network Scheduling on Multiple Hybrid Memory Cube-based Processing-in-Memory
Young Sik Lee, Tae Hee Han
IEEE Access, Vol. 9, pp. 68561-68572, May 2021.
#Memory-Centric #AI Accel #PIM #DNN Scheduling Abstract