🚀 Accelerating Generative AI

ASIC

C JNPU: A 1.04TFLOPS Joint-DNN Training Processor with Speculative Quantization and Triple Heterogeneity on Microarchitecture / Precision / Dataflow [paper]

IEEE European Solid-State Circuits Conference (ESSCIRC), 2023

Je Yang, Sukbin Lim, Sukjin Lee, Jae-Young Kim, and Joo-Young Kim

FPGA

C FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism [paper]

ACM/IEEE Design Automation Conference (DAC), 2021

Je Yang, Seongmin Hong, and Joo-Young Kim

C LearningGroup: A Real-Time Sparse Training on FPGA via Learnable Weight Grouping for Multi-Agent Reinforcement Learning [paper]

IEEE International Conference on Field Programmable Technology (ICFPT), 2022

Je Yang, JaeUk Kim, and Joo-Young Kim

C ReconFormer: A Multi-Level Run-time Reconfigurable System-on-Chip for Accelerating Transformers [paper]

IEEE International Conference on Field-Programmable Logic and Applications (FPL), 2025

Je Yang, Gabriele Tombesi, Joseph Zuckerman, and Luca P. Carloni

LanguageModel

C ReconFormer: A Multi-Level Run-time Reconfigurable System-on-Chip for Accelerating Transformers [paper]

IEEE International Conference on Field-Programmable Logic and Applications (FPL), 2025

Je Yang, Gabriele Tombesi, Joseph Zuckerman, and Luca P. Carloni

MultiTenant-CNN

C JNPU: A 1.04TFLOPS Joint-DNN Training Processor with Speculative Quantization and Triple Heterogeneity on Microarchitecture / Precision / Dataflow [paper]

IEEE European Solid-State Circuits Conference (ESSCIRC), 2023

Je Yang, Sukbin Lim, Sukjin Lee, Jae-Young Kim, and Joo-Young Kim

Pruning

C LearningGroup: A Real-Time Sparse Training on FPGA via Learnable Weight Grouping for Multi-Agent Reinforcement Learning [paper]

IEEE International Conference on Field Programmable Technology (ICFPT), 2022

Je Yang, JaeUk Kim, and Joo-Young Kim

Quantization

C FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism [paper]

ACM/IEEE Design Automation Conference (DAC), 2021

Je Yang, Seongmin Hong, and Joo-Young Kim

C JNPU: A 1.04TFLOPS Joint-DNN Training Processor with Speculative Quantization and Triple Heterogeneity on Microarchitecture / Precision / Dataflow [paper]

IEEE European Solid-State Circuits Conference (ESSCIRC), 2023

Je Yang, Sukbin Lim, Sukjin Lee, Jae-Young Kim, and Joo-Young Kim

ReinforcementLearning

C FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism [paper]

ACM/IEEE Design Automation Conference (DAC), 2021

Je Yang, Seongmin Hong, and Joo-Young Kim

C LearningGroup: A Real-Time Sparse Training on FPGA via Learnable Weight Grouping for Multi-Agent Reinforcement Learning [paper]

IEEE International Conference on Field Programmable Technology (ICFPT), 2022

Je Yang, JaeUk Kim, and Joo-Young Kim

Jenna Yang

ASIC

C JNPU: A 1.04TFLOPS Joint-DNN Training Processor with Speculative Quantization and Triple Heterogeneity on Microarchitecture / Precision / Dataflow [paper]

FPGA

C FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism [paper]

C LearningGroup: A Real-Time Sparse Training on FPGA via Learnable Weight Grouping for Multi-Agent Reinforcement Learning [paper]

C ReconFormer: A Multi-Level Run-time Reconfigurable System-on-Chip for Accelerating Transformers [paper]

LanguageModel

C ReconFormer: A Multi-Level Run-time Reconfigurable System-on-Chip for Accelerating Transformers [paper]

MultiTenant-CNN

C JNPU: A 1.04TFLOPS Joint-DNN Training Processor with Speculative Quantization and Triple Heterogeneity on Microarchitecture / Precision / Dataflow [paper]

Pruning

C LearningGroup: A Real-Time Sparse Training on FPGA via Learnable Weight Grouping for Multi-Agent Reinforcement Learning [paper]

Quantization

C FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism [paper]

C JNPU: A 1.04TFLOPS Joint-DNN Training Processor with Speculative Quantization and Triple Heterogeneity on Microarchitecture / Precision / Dataflow [paper]

ReinforcementLearning

C FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism [paper]

C LearningGroup: A Real-Time Sparse Training on FPGA via Learnable Weight Grouping for Multi-Agent Reinforcement Learning [paper]