http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Kim, Wansun,Choi, Joonhyeong,Kim, Jae-Han,Kim, Taesu,Lee, Changyeon,Lee, Seungjin,Kim, Mingoo,Kim, Bumjoon J.,Kim, Taek-Soo American Chemical Society 2018 Chemistry of materials Vol.30 No.6
<P>High fracture resistance of polymer solar cells (PSCs) is of great importance to ensure long-term mechanical reliability, especially considering their potential in roll-to-roll printing processes and flexible devices. In this paper, we compare mechanical properties, such as the cohesive fracture energy, elastic modulus, and crack-onset strain, of all-polymer solar cells (all-PSCs) and fullerene-based solar cells (PCBM-PSCs) based on the same, representative low-bandgap polymer donor (PTB7-Th) as a function of acceptor content. The all-PSCs exhibit higher fracture energy (2.45 J m<SUP>-2</SUP>) than PCBM-PSCs (0.29 J m<SUP>-2</SUP>) at optimized device conditions. Additionally, a 15-fold higher crack-onset strain is observed in all-PSCs than in PCBM-PSCs. Dramatically different mechanical compliances observed for all-PSCs and PCBM-PSCs are investigated in detail by analysis of the blend morphologies as a function of acceptor content (either P(NDI2HD-T) or PCBM acceptors). The superior fracture resistance of all-PSCs is attributed to the more ductile characteristics of the polymer acceptor and the large degree of plastic deformation during crack growth, in contrast to the brittle nature of PCBM and the weak interaction between the polymer-rich phase and highly aggregated PCBM-rich domains. Therefore, this work demonstrates that replacing a small-molecule acceptor (i.e., PCBM) with polymeric materials can be an effective strategy toward mechanically robust PSCs.</P> [FIG OMISSION]</BR>
Input Voltage Mapping Optimized for Resistive Memory-Based Deep Neural Network Hardware
Taesu Kim,Hyungjun Kim,Jinseok Kim,Jae-Joon Kim IEEE 2017 IEEE electron device letters Vol.38 No.9
<P>Artificial neural network (ANN) computations based on graphics processing units (GPUs) consume high power. Resistive random-access memory (RRAM) has been gaining attention as a promising technology for implementing power-efficient ANNs, replacing GPU. However, nonlinear I-V characteristics of RRAM devices have been limiting its use for ANN implementation. In this letter, we propose a method and a circuit to address issues due to the nonlinear I-V characteristics. We demonstrate the feasibility of the method by simulating its application to multiple neural networks, from multi-layer perceptron to deep convolutional neural network based on a typical RRAM model. Results from classifying datasets including ImageNet show that the proposed method produces much higher accuracy than the naive linear mapping for a wide range of nonlinearity.</P>
Kim, Taesu,Choi, Joonhyeong,Kim, Hyeong Jun,Lee, Wonho,Kim, Bumjoon J. American Chemical Society 2017 Macromolecules Vol.50 No.17
<P>We compared the thermal and morphological stability of all-polymer solar cells (all-PSCs) and fullerene-based PSCs (fullerene-PSCs) having the same polymer donor (PBDTTTPD), which provided comparable peak power conversion efficiencies (PCEs) of >6%. We observed a remarkable contrast in thermal stability dependent upon the acceptor composition in the active layer, with the performance of the fullerene-PSCs completely deteriorating after annealing for 5 h at 150 °C, whereas that of the all-PSCs remained stable even after annealing for 50 h at 150 °C. Pronounced phase separation was observed in the active layer of the fullerene-PSCs at two different length scales. In stark contrast, almost no morphological changes in the all-PSCs were observed, likely due to the low diffusion kinetics of the polymer acceptors. To develop a comprehensive understanding of the role of polymer acceptor on the thermal stability of devices, the morphology of ternary blend active layers composed of PBDTTTPD:polymer acceptor:fullerene acceptor with different fullerene contents was examined while annealing at 150 °C. The ternary blends showed two extreme trends of all-PSC- and fullerene-PSC-like behavior in thermal stability depending on the PCBM content. When included in the active layer as <30 wt % of the acceptor mixture, fullerene was well-dispersed in the amorphous portion of the donor/acceptor polymer blend under thermal stress and led to thermally stable devices with a higher PCE (7.12%) than both all-PSCs without fullerene (6.67%) and polymer-fullerene active layers without a polymeric acceptor (6.12%).</P> [FIG OMISSION]</BR>
Hi, KIA! 기계 학습을 이용한 기동어 기반 감성 분류
김태수 ( Taesu Kim ),김영우 ( Yeongwoo Kim ),김근형 ( Keunhyeong Kim ),김철민 ( Chul Min Kim ),전형석 ( Hyung Seok Jun ),석현정 ( Hyeon-jeong Suk ) 한국감성과학회 2021 감성과학 Vol.24 No.1
본 연구에서는 승용차에서 사람들이 기기를 사용하기 위해 사용하는 기동어인 “Hi, KIA!”의 감성을 기계학습을 기반으로 분류가 가능한가에 대해 탐색하였다. 감성 분류를 위해 신남, 화남, 절망, 보통 총 4가지 감정별로 3가지 시나리오를 작성하여, 자동차 운전 상황에서 발생할 수 있는 12가지의 사용자 감정 시나리오를 제작하였다. 시각화 자료를 기반으로 총 9명의 대학생을 대상으로 녹음을 진행하였다. 수집된 녹음 파일의 전체 문장에서 기동어 부분만 별도로 추출하는 과정을 거쳐, 전체 문장 파일, 기동어 파일 총 두 개의 데이터 세트로 정리되었다. 음성 분석에서는 음향 특성을 추출하고 추출된 데이터를 svmRadial 방법을 이용하여 기계 학습 기반의 알고리즘을 제작해, 제작된 알고리즘의 감정 예측 정확성 및 가능성을 파악하였다. 9명의 참여자와 4개의 감정 카테고리를 통틀어 기동어의 정확성(60.19%: 22~81%)과 전체 문장의 정확성(41.51%)을 비교했다. 또한, 참여자 개별로 정확도와 민감도를 확인하였을 때, 성능을 보임을 확인하였으며, 각 사용자 별 기계 학습을 위해 선정된 피쳐들이 유사함을 확인하였다. 본 연구는 기동어만으로도 사용자의 감정 추출과 보이스 인터페이스 개발 시 기동어 감정 파악 기술이 잠재적으로 적용 가능한데 대한 실험적 증거를 제공할 수 있을 것으로 기대한다. This study explored users’ emotional states identified from the wake-up words ―“Hi, KIA!”―using a machine learning algorithm considering the user interface of passenger cars’ voice. We targeted four emotional states, namely, excited, angry, desperate, and neutral, and created a total of 12 emotional scenarios in the context of car driving. Nine college students participated and recorded sentences as guided in the visualized scenario. The wake-up words were extracted from whole sentences, resulting in two data sets. We used the soundgen package and svmRadial method of caret package in open source-based R code to collect acoustic features of the recorded voices and performed machine learning-based analysis to determine the predictability of the modeled algorithm. We compared the accuracy of wake-up words (60.19%: 22%~81%) with that of whole sentences (41.51%) for all nine participants in relation to the four emotional categories. Accuracy and sensitivity performance of individual differences were noticeable, while the selected features were relatively constant. This study provides empirical evidence regarding the potential application of the wake-up words in the practice of emotion-driven user experience in communication between users and the artificial intelligence system.
Lee, Wonho,Kim, Jae-Han,Kim, Taesu,Kim, Seonha,Lee, Changyeon,Kim, Jin-Seong,Ahn, Hyungju,Kim, Taek-Soo,Kim, Bumjoon J. The Royal Society of Chemistry 2018 Journal of Materials Chemistry A Vol.6 No.10
<P>In this study, we demonstrate that the introduction of small amounts of phenyl-C71-butyric acid methyl ester (PC71BM) into all-polymer solar cells (all-PSCs) increases the photovoltaic performance without compromising mechanical properties. Ternary blend polymer solar cells (ternary-PSCs) consisting of a polymer donor (PTB7-Th) and an acceptor mixture with different weight ratios of a polymeric acceptor (P(NDI2HD-T2)) and PC71BM demonstrate the effects of PC71BM loading on the power conversion efficiency (PCE) and mechanical properties. A significant enhancement in the PCEs of ternary-PSCs, from 6.32% to 7.33%, is observed when PC71BM is added into the active layer as up to 30 wt% of the acceptor mixture. Importantly, the excellent mechanical properties (<I>i.e.</I>, crack onset strain = 11.6%, toughness = 2237 J m<SUP>−3</SUP>) of the blend films are well preserved at PC71BM loadings at or below 30 wt%. In contrast, both the PCE and the mechanical performance of the ternary-PSCs significantly decrease at higher PC71BM loadings (>50 wt%). Detailed morphological analysis<I>via</I>grazing incidence X-ray scattering measurements reveals that PC71BM molecules are well-dispersed in the amorphous portion of the active layer at PC71BM loadings up to 30 wt%. Therefore, both the mechanical and photovoltaic performances of the ternary-PSCs correlate closely with their morphological behavior, particularly in terms of the mixing behavior of PC71BM with polymers. The well-dispersed PC71BM molecules in the amorphous polymer domains facilitate efficient exciton dissociation, whereas the formation of PC71BM aggregates above a critical concentration causes severe mechanical degradation of the ternary-PSCs due to the presence of weak interfaces between the brittle PC71BM and polymer domains. Therefore, the ternary blends with optimal content of polymer/fullerene acceptors represent important candidates for flexible and wearable solar cells that require both high mechanical and photovoltaic performances.</P>
Efficient Synapse Memory Structure for Reconfigurable Digital Neuromorphic Hardware
Kim, Jinseok,Koo, Jongeun,Kim, Taesu,Kim, Jae-Joon Frontiers Media S.A. 2018 Frontiers in neuroscience Vol.12 No.-
<P>Spiking Neural Networks (SNNs) have high potential to process information efficiently with binary spikes and time delay information. Recently, dedicated SNN hardware accelerators with on-chip synapse memory array are gaining interest in overcoming the limitations of running software-based SNN in conventional Von Neumann machines. In this paper, we proposed an efficient synapse memory structure to reduce the amount of hardware resource usage while maintaining performance and network size. In the proposed design, synapse memory size can be reduced by applying presynaptic weight scaling. In addition, axonal/neuronal offsets are applied to implement multiple layers on a single memory array. Finally, a transposable memory addressing scheme is presented for faster operation of spike-timing-dependent plasticity (STDP) learning. We implemented a SNN ASIC chip based on the proposed scheme with 65 nm CMOS technology. Chip measurement results showed that the proposed design provided up to 200X speedup over CPU while consuming 53 mW at 100 MHz with the energy efficiency of 15.2 pJ/SOP.</P>
Photopolymer-based demultiplexers with Superposed holographic gratings
Taesu Kim,Seunghwan Chung,Seunghoon Han,Byoungho Lee IEEE 2005 IEEE photonics technology letters Vol.17 No.3
<P>A photopolymer-based demultiplexer structure with two superposed holographic gratings is proposed and demonstrated for the first time. The demultiplexer is constructed from a holographic recording with two coherent beam sets. In the proposed scheme, two different channels of optical signals are coupled into a common single-mode fiber (SMF) by means of diffractions from the two superposed gratings. The coupled spectral channels are selected by tuning the spatial location of the SMF. The wavelength separation between two spectral channels that couple into the same optical fiber can be controlled by varying the angle between the two beams used for the formation of the holographic gratings.</P>