Kwantae Kim

Kwantae Kim

Auto-generated profile, based on publicly available information

Kwantae Kim

Department of Electronics and Nanoengineering, School of Electrical Engineering, Aalto University, Espoo, Finland

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 August 2022 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Open access
KKKwantae Kim
Kwantae Kim
CGChang Gao
Chang Gao
RGRui Graca
Rui Graca
IKIlya Kiselev
Ilya Kiselev
HYHoi-Jun Yoo
Hoi-Jun Yoo
SLShih-Chii Liu
Shih-Chii Liu
This article presents the first keyword spotting (KWS) IC that uses a ring-oscillator-based time-domain processing technique for its analog feature extractor (FEx). Its extensive usage of time-encoding schemes allows the analog audio signal to be processed in a fully time-domain manner except for the voltage-to-time conversion stage of the analog front end. Benefiting from fundamental building blocks based on digital logic gates, it offers better technology scalability compared to conventional voltage-domain designs. Fabricated in a 65-nm CMOS process, the prototyped KWS IC occupies 2.03 mm2 and dissipates 23- $\mu$ W power consumption, including analog FEx and digital neural network classifier. The 16-channel time-domain FEx achieves a 54.89-dB dynamic range for 16-ms frame shift size while consuming 9.3 $\mu$ W. The measurement result verifies that the proposed IC performs a 12-class KWS task on the Google Speech Command dataset (GSCD) with $>$ 86% accuracy and 12.4-ms latency.

19 May 2023 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
SKSangyeob Kim
Sangyeob Kim
SKSangjin Kim
Sangjin Kim
SUSoyeon Um
Soyeon Um
SKSoyeon Kim
Soyeon Kim
KKKwantae Kim
Kwantae Kim
HYHoi-Jun Yoo
Hoi-Jun Yoo
A highly energy-efficient neuromorphic computing-in-memory (Neuro-CIM) processor is proposed for ultralow-power deep learning applications. Neuro-CIM can support spiking neural network (SNN) to eliminate the power and area overhead of previous CIM processor. The sign extended bits gating reduces the bitline (BL) voltage switching rate due to negative small-magnitude weights allowing 38% power reduction at 8-b weight condition and 25% at 4-b weight condition. In addition, Neuro-CIM replaces high-precision analog-to-digital converter (ADC) with 1-b comparator by exploiting the characteristic of the SNN, and thus, power and area efficiencies are significantly enhanced. Furthermore, the early stopping scheme terminates unnecessary neuronal operations, reducing power consumption by 31%. In addition, the analog and digital networks are integrated for high reconfigurability and energy efficiency. The analog network with voltage folding circuit enables accurate analog-domain aggregation by increasing the dynamic range without compromising the voltage resolution. The digital network-in-memory supports input–output channel extension for high reconfigurability and input data reuse scheme for reducing input memory (IMEM) access. Neuro-CIM is fabricated in 28-nm CMOS technology and occupies the 2.9-mm $^{2}$ die area. It achieves the state-of-the-art energy consumption per classification of 0.72 $\mu$ J and 92.1% accuracy for CIFAR-10 with 4-b input and 4-b weight and 372.2 $\mu$ J and 65.8% accuracy for ImageNet with 6-b input and 8-b weight at 200 MHz, and 1.1-V conditions. Moreover, the proposed CIM processor achieves 310.37 tera operations per second/watt (TOPS/W) and 90.7% accuracy with 4-b input and 1-b weight for Canadian Institute for Advanced Research, 10 classes (CIFAR-10) classification.

1 August 2020 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
KKKwantae Kim
Kwantae Kim
JKJi-Hoon Kim
Ji-Hoon Kim
SGSurin Gweon
Surin Gweon
MKMinseo Kim
Minseo Kim
HYHoi-Jun Yoo
Hoi-Jun Yoo
Continuous monitoring of fluid status through bio-impedance (Bio-Z) measurement of thoracic magnitude or phase is critical in reducing the mortality of chronic heart failure (CHF) patients. However, the stringent power constraints of implantable devices force the design of Bio-Z sensors to be highly challenging. We present a sub-10-μW Bio-Z sensor IC operating from a 0.5 V-supply, sustaining a sub-1° of phase error even under 10% of supply and 0 °C-70 °C of temperature variations. Fabricated in a 65-nm CMOS process, this sensor IC exhibits a 15.28-mΩ/√Hz of input-referred impedance noise performance, occupying a 4.83 mm². Its linearity performance shows that the magnitude and the phase of Bio-Z can be measured within the error of 2% and 0.4°, respectively, given that the magnitude of load impedance is less than 126 Ω, and thus, it is suitable for fluid status monitoring applications. Experimental measurement on the human chest demonstrates its capabilities of thoracic impedance variance (TIV) and impedance cardiography (ICG) monitoring.

6 August 2021 in IEEE Journal of Solid-State Circuits
Verified
Journal Article
Closed access
KKKwantae Kim
Kwantae Kim
SKSangyeob Kim
Sangyeob Kim
HYHoi-Jun Yoo
Hoi-Jun Yoo
This article presents a low-power, low-distortion, and compact mixed-signal sinusoidal current generator (CG) IC for bio-impedance (Bio-Z) sensing applications. By utilizing the digital $\Delta \Sigma $ modulation to bridge the digitally synthesized sinewave data and the analog-domain voltage output, implementation of a low-distortion sinewave lookup table (LUT) is significantly simplified. Not only does this approach lead to the reduced number of routing wires, pin allocation, and output registers after logic synthesis but also, it allows the use of dynamic element matching (DEM) with a low-cost, before interfacing with the digital-to-analog converter (DAC). In addition, 0.5 V of low-supply voltage exploits a highly power-efficient operation for both synthesized digital logic and its subsequent analog circuit chain. To suppress the noise-floor increase that is induced from the reset signal of capacitor-DAC, a half-period reset scheme is introduced. The prototype chip is fabricated in 65-nm CMOS technology and it outputs $2~\mu \text{A}_{\text {PP}}$ of amplitude with 20 kHz of sinusoidal frequency. It records the first sinusoidal CG IC that achieves sub- $10~\mu \text{W}$ ( $6.2~\mu \text{W}$ ) of power consumption and sub-0.1% (0.088%) of total harmonic distortion (THD) at the same time which has not been demonstrated ever in the prior arts. Furthermore, 0.059 mm 2 of its compact area facilitates the low-cost production of Bio-Z sensor systems.

21 July 2023 in IEEE Circuits and Systems Magazine
Verified
Journal Article
Closed access
KKKwantae Kim
Kwantae Kim
SLShih-Chii Liu
Shih-Chii Liu
Edge audio devices can reduce data bandwidth requirements by pre-processing input speech on the device before transmission to the cloud. As edge devices are required to ensure always-on operation, their stringent power constraints pose several design challenges and force IC designers to look for solutions that use low standby power. One promising bio-inspired approach is to combine the continuous-time analog filter channels with a small memory footprint deep neural network that is trained on edge tasks such as keyword spotting, thereby allowing all blocks to be embedded in an IC. This paper reviews the historical background of the continuous-time analog filter circuits that have been used as feature extractors for current edge audio devices. Starting from the interpretation of a basic biquad filter as a two-integrator-loop topology, we introduce the progression in the design of secondorder low-pass and band-pass filters ranging from OTA-based to source-follower-based architectures. We also derive and analyze the small-signal transfer function and discuss their usage in edge audio applications.

28 November 2024 in IEEE Transactions on Circuits and Systems for Artificial Intelligence
Verified
Journal Article
Closed access
QCQinyu Chen
Qinyu Chen
KKKwantae Kim
Kwantae Kim
CGChang Gao
Chang Gao
SZSheng Zhou
Sheng Zhou
TJTaekwnag Jang
Taekwnag Jang
SLShih-Chii Liu
Shih-Chii Liu
This paper introduces DeltaKWS, to the best of our knowledge, the first ΔRNN-enabled fine-grained temporal sparsity-aware Keyword Spotting (KWS) integrated circuit (IC) for voice-controlled devices. The 65nm prototype chip features a number of techniques to enhance performance, area, and power efficiencies, specifically: 1) a bio-inspired delta-gated recurrent neural network (ΔRNN) classifier leveraging temporal similarities between neighboring feature vectors extracted from input frames and network hidden states, eliminating unnecessary operations and memory accesses; 2) an infinite impulse response (IIR) bandpass filter (BPF)-based feature extractor (FEx) that leverages mixed-precision quantization, low-cost computing structure and channel selection; 3) a 24 kB 0.6V near- V TH weight static random-access memory (SRAM) that achieves 6.6× lower read power than the foundry-provided SRAM. From chip measurement results, we show that the DeltaKWS achieves an 11/12-class Google Speech Command Dataset (GSCD) accuracy of 90.5%/89.5% respectively and energy consumption of 36 nJ/decision in 65nm CMOS process. At 87% temporal sparsity, computing latency and energy/inference are reduced by 2.4×/3.4×, respectively. The IIR BPF-based FEx, ΔRNN accelerator, and 24 kB near- V TH SRAM blocks occupy 0.084mm 2 , 0.319mm 2 , and 0.381mm 2 respectively (0.78mm 2 in total).