Jung Nam Kim
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jung Nam Kim has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Yoon Kim
University of Seoul
5 shared publications - Dae Hwan Kim
Kookmin University
2 shared publications - Minsuk Koo
University of Seoul
2 shared publications - Chan-Gi Yook
Department of Electrical and Information Engineering, Seoul National University of Science and Technology, Seoul, Republic of Korea
1 shared publication - Changhyun Kim
Pohang University of Science and Technology (POSTECH), Pohang, South Korea
1 shared publication - Changwook Kim
School of Electrical Engineering, Kookmin University, Seoul, South Korea
1 shared publication
Latest Publications
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 16 (2) , 218-226
Voltage-Summation-Based Compute-in-Memory SRAM Macro With Input-Bit and Synaptic-Bit Slicing
Compute-in-memory (CIM) technology has emerged as a promising solution for energy-efficient matrix-vector multiplication (MVM), addressing the inherent limitations of von Neumann architecture. However, conventional Kirchhoff’s current law based CIM technologies suffer from drawbacks such as I-R drop and energy inefficiency due to static current flow over a wide dynamic range. This paper introduces an energy-efficient and I-R drop-free CIM architecture based on capacitive coupling principle. Synaptic weights represented in 4-bit two’s complement format are stored in SRAM-based 9T1C bit-cells. The architecture utilizes input-bit slicing method to perform MVM operations without the need for multi-bit digital-to-analog converters (DACs), significantly improving energy efficiency. A neuron circuit employs successive integration and rescaling (SIR) process to handle input-bit slicing encoded analog MVM results efficiently. The fabricated 16 kb, 28-nm prototype chip achieves an energy efficiency of 230 TOPS/W and a throughput of 32.19 GOPS for MVM operation with 4-bit resolution input and synaptic weights at a 100 MHz clock frequency. Even with ADC-induced analog variability, the proposed architecture achieves 90.64% and 97.99% inference accuracy for single-layer and multi-layer networks on the MNIST dataset, and 86.25% on CIFAR-10, validating its effectiveness and practical viability for low-power edge inference in complex workloads.
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 16 (2) , 218-226
Voltage-Summation-Based Compute-in-Memory SRAM Macro With Input-Bit and Synaptic-Bit Slicing
Compute-in-memory (CIM) technology has emerged as a promising solution for energy-efficient matrix-vector multiplication (MVM), addressing the inherent limitations of von Neumann architecture. However, conventional Kirchhoff’s current law based CIM technologies suffer from drawbacks such as I-R drop and energy inefficiency due to static current flow over a wide dynamic range. This paper introduces an energy-efficient and I-R drop-free CIM architecture based on capacitive coupling principle. Synaptic weights represented in 4-bit two’s complement format are stored in SRAM-based 9T1C bit-cells. The architecture utilizes input-bit slicing method to perform MVM operations without the need for multi-bit digital-to-analog converters (DACs), significantly improving energy efficiency. A neuron circuit employs successive integration and rescaling (SIR) process to handle input-bit slicing encoded analog MVM results efficiently. The fabricated 16 kb, 28-nm prototype chip achieves an energy efficiency of 230 TOPS/W and a throughput of 32.19 GOPS for MVM operation with 4-bit resolution input and synaptic weights at a 100 MHz clock frequency. Even with ADC-induced analog variability, the proposed architecture achieves 90.64% and 97.99% inference accuracy for single-layer and multi-layer networks on the MNIST dataset, and 86.25% on CIFAR-10, validating its effectiveness and practical viability for low-power edge inference in complex workloads.
- Open Access
- 25 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 35894-35900
ES-SNN: e-Flash Memory-Based Stochastic Spiking Neural Networks for Edge Inference
In this work, we propose an e-Flash memory-based Stochastic Spiking Neural Network (ES-SNN), which integrates stochastic spiking neurons with e-Flash memory-based multi-level synapses for probabilistic computation. We design an energy-efficient, capacitor-less stochastic output neuron using a 28 nm CMOS process. The e-Flash-based non-volatile synapses achieve 3-bit weight storage per cell, enabling compact and energy-efficient inference. System level evaluation results demonstrate that the proposed ES-SNN achieves an energy efficiency of 608 Tera Operations per Second per Watt (TOPS/W) for a single-layer fully connected SNN, highlighting its potential for low-power neuromorphic computing applications.
- 25 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 35894-35900
Open AccessES-SNN: e-Flash Memory-Based Stochastic Spiking Neural Networks for Edge Inference
In this work, we propose an e-Flash memory-based Stochastic Spiking Neural Network (ES-SNN), which integrates stochastic spiking neurons with e-Flash memory-based multi-level synapses for probabilistic computation. We design an energy-efficient, capacitor-less stochastic output neuron using a 28 nm CMOS process. The e-Flash-based non-volatile synapses achieve 3-bit weight storage per cell, enabling compact and energy-efficient inference. System level evaluation results demonstrate that the proposed ES-SNN achieves an energy efficiency of 608 Tera Operations per Second per Watt (TOPS/W) for a single-layer fully connected SNN, highlighting its potential for low-power neuromorphic computing applications.
- Open Access
- 22 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 23881-23886
Current-Voltage Modeling of DRAM Cell Transistor Using Genetic Algorithm and Deep Learning
Accurate current-voltage ( I-V ) modeling based on the Berkeley short-channel insulated-gate field-effect transistor model (BSIM) is pivotal for integrated circuit simulation. However, the current BSIM model does not support a buried-channel-array transistor (BCAT), which is the structure of the state-of-the-art commercial dynamic random access memory (DRAM) cell transistor. In this work, we propose an intelligent I-V modeling technique that combines genetic algorithm (GA) and deep learning (DL). This hybrid technique facilitates both optimization of BSIM parameter and accurate I-V modeling, even for devices not originally supported by BSIM. Additionally, we extended application of the DL to model one of the principal degradation mechanisms of transistor, the hot-carrier degradation (HCD). The successful modeling results of I-V characteristic and device degradation demonstrated that devices not supported by BSIM can be accurately modeled for integrated circuit simulations.
- 22 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 23881-23886
Open AccessCurrent-Voltage Modeling of DRAM Cell Transistor Using Genetic Algorithm and Deep Learning
Accurate current-voltage ( I-V ) modeling based on the Berkeley short-channel insulated-gate field-effect transistor model (BSIM) is pivotal for integrated circuit simulation. However, the current BSIM model does not support a buried-channel-array transistor (BCAT), which is the structure of the state-of-the-art commercial dynamic random access memory (DRAM) cell transistor. In this work, we propose an intelligent I-V modeling technique that combines genetic algorithm (GA) and deep learning (DL). This hybrid technique facilitates both optimization of BSIM parameter and accurate I-V modeling, even for devices not originally supported by BSIM. Additionally, we extended application of the DL to model one of the principal degradation mechanisms of transistor, the hot-carrier degradation (HCD). The successful modeling results of I-V characteristic and device degradation demonstrated that devices not supported by BSIM can be accurately modeled for integrated circuit simulations.
- Open Access
- 7 September 2023
- Published by MDPI AG in Micromachines
- Vol. 14 (9) , 1753
Design Strategies of 40 nm Split-Gate NOR Flash Memory Device for Low-Power Compute-in-Memory Applications
The existing von Neumann architecture for artificial intelligence (AI) computations suffers from excessive power consumption and memory bottlenecks. As an alternative, compute-in-memory (CIM) technology has been emerging. Among various CIM device candidates, split-gate NOR flash offers advantages such as a high density and low on-state current, enabling low-power operation, and benefiting from a high level of technological maturity. To achieve high energy efficiency and high accuracy in CIM inference chips, it is necessary to optimize device design by targeting low power consumption at the device level and surpassing baseline accuracy at the system level. In split-gate NOR flash, significant factors that can cause CIM inference accuracy drop are the device conductance variation, caused by floating gate charge variation, and a low on-off current ratio. Conductance variation generally has a trade-off relationship with the on-current, which greatly affects CIM dynamic power consumption. In this paper, we propose strategies for designing optimal devices by adjusting oxide thickness and other structural parameters. As a result of setting Tox,FG to 13.4 nm, TIPO to 4.6 nm and setting other parameters to optimal points, the design achieves erase on-current below 2 μA, program on-current below 10 pA, and off-current below 1 pA, while maintaining an inference accuracy of over 92%.
- 7 September 2023
- Published by MDPI AG in Micromachines
- Vol. 14 (9) , 1753
Open AccessDesign Strategies of 40 nm Split-Gate NOR Flash Memory Device for Low-Power Compute-in-Memory Applications
The existing von Neumann architecture for artificial intelligence (AI) computations suffers from excessive power consumption and memory bottlenecks. As an alternative, compute-in-memory (CIM) technology has been emerging. Among various CIM device candidates, split-gate NOR flash offers advantages such as a high density and low on-state current, enabling low-power operation, and benefiting from a high level of technological maturity. To achieve high energy efficiency and high accuracy in CIM inference chips, it is necessary to optimize device design by targeting low power consumption at the device level and surpassing baseline accuracy at the system level. In split-gate NOR flash, significant factors that can cause CIM inference accuracy drop are the device conductance variation, caused by floating gate charge variation, and a low on-off current ratio. Conductance variation generally has a trade-off relationship with the on-current, which greatly affects CIM dynamic power consumption. In this paper, we propose strategies for designing optimal devices by adjusting oxide thickness and other structural parameters. As a result of setting Tox,FG to 13.4 nm, TIPO to 4.6 nm and setting other parameters to optimal points, the design achieves erase on-current below 2 μA, program on-current below 10 pA, and off-current below 1 pA, while maintaining an inference accuracy of over 92%.
- Open Access
- 27 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 20196-20201
Short- and Long-Term Memory Based on a Floating-Gate IGZO Synaptic Transistor
Short- and long-term neuroplasticity behaviors are key mechanisms for various activities. In this paper, we propose a synaptic transistor with a floating-gate (FG) node and an amorphous InGaZnO (IGZO) channel layer. The proposed device emulates the neuroplasticity functions of both short-term memory (STM) and long-term memory (LTM) through the control of the amplitude and the number of input pulses. The STM operated by ion movement in the gate dielectrics occurs when the input amplitude is relatively small (< 9 V). The LTM operated through the storage of electrons in the FG occurs when the input amplitude is relatively large (>10 V). In addition, as the number of input pulses increases, information is stored for a longer time. Our FG IGZO synaptic transistor could be a promising device solution for brain-inspired computing systems.
- 27 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 20196-20201
Open AccessShort- and Long-Term Memory Based on a Floating-Gate IGZO Synaptic Transistor
Short- and long-term neuroplasticity behaviors are key mechanisms for various activities. In this paper, we propose a synaptic transistor with a floating-gate (FG) node and an amorphous InGaZnO (IGZO) channel layer. The proposed device emulates the neuroplasticity functions of both short-term memory (STM) and long-term memory (LTM) through the control of the amplitude and the number of input pulses. The STM operated by ion movement in the gate dielectrics occurs when the input amplitude is relatively small (< 9 V). The LTM operated through the storage of electrons in the FG occurs when the input amplitude is relatively large (>10 V). In addition, as the number of input pulses increases, information is stored for a longer time. Our FG IGZO synaptic transistor could be a promising device solution for brain-inspired computing systems.