• 제목/요약/키워드: stacked

검색결과 1,132건 처리시간 0.028초

CNN 추론 연산 가속기를 위한 곱셈기 최적화 설계 (Design of Multipliers Optimized for CNN Inference Accelerators)

  • 이재우;이재성
    • 한국정보통신학회논문지
    • /
    • 제25권10호
    • /
    • pp.1403-1408
    • /
    • 2021
  • AI 프로세서를 FPGA 기반으로 구현하는 연구가 최근 활발하게 진행되고 있다. Deep Convolutional Neural Networks (CNN) 는 AI 프로세서가 수행하는 기본적인 연산 구조로서 매우 방대한 양의 곱셈을 필요로 한다. CNN 추론 연산에서 사용되는 곱셈 계수는 상수라는 점과 FPGA 은 특정 계수에 맞춰진 곱셈기 설계가 용이하다는 점에 착안하여 곱셈기를 최적화 구현할 수 있는 방법을 제안한다. 본 방법은 2의 보수와 분배법칙을 활용하여 곱셈 계수에서 값이 1인 비트의 개수를 최소화하여 필요한 적층 덧셈기의 개수를 절감한다. CNN 을 FPGA 에 구현한 실제 예제에 본 방법을 적용해본 결과 로직 사용량은 최대 30.2%까지, 신호 전달 지연은 최대 22%까지 줄어들었다. ASIC 전용 칩으로 구현할 경우에도 하드웨어 면적은 최대 35%까지, 신호 전달 지연은 최대 19.2%까지 줄어드는 것으로 나타났다.

Josephson effect of the superconducting van der Waals junction

  • Park, Sungyu;Kwon, Chang Il;Kim, Jun Sung
    • 한국초전도ㆍ저온공학회논문지
    • /
    • 제23권2호
    • /
    • pp.6-9
    • /
    • 2021
  • 단결정의 덩어리 시료로부터 박리된 NbSe2와 FeSe 박막을 이용한 동종 초전도 vdW 접합을 만들고 저온전도측정을 통해 JJ 효과를 확인하였다. 각 물질의 초전도 틈을 측정하였으며 기존 결과와 잘 일치하는 것을 확인하였다. FeSe 접합의 경우 NbSe2 접합과 다르게 JJ가 구현되지 않는 경우가 다수 발생하는데 이는 NbSe2와 구분되는 FeSe의 특성때문으로 보이며 일관성있는 JJ 구현을 위해서 저온 접합이나 접합 각도 의존성을 통한 향후 연구가 필요하다. 그럼에도 불구하고, 본 연구 결과는 FeSe 2차원 결정을 이용해 JJ가 잘 구현될 수 있음을 실험적으로 보인 첫 사례로서 그 의의를 갖는다.

Application of black phosphorus nanodots to live cell imaging

  • Shin, Yong Cheol;Song, Su-Jin;Lee, Yu Bin;Kang, Moon Sung;Lee, Hyun Uk;Oh, Jin-Woo;Han, Dong-Wook
    • 생체재료학회지
    • /
    • 제22권4호
    • /
    • pp.352-359
    • /
    • 2018
  • Background: Black phosphorus (BP) has emerged as a novel class of nanomaterials owing to its unique optical and electronic properties. BP, a two-dimensional (2D) nanomaterial, is a structure where phosphorenes are stacked together in layers by van der Waals interactions. However, although BP nanodots have many advantages, their biosafety and biological effect have not yet been elucidated as compared to the other nanomaterials. Therefore, it is particularly important to assess the cytotoxicity of BP nanodots for exploring their potentials as novel biomaterials. Methods: BP nanodots were prepared by exfoliation with a modified ultrasonication-assisted solution method. The physicochemical properties of BP nanodots were characterized by transmission electron microscopy, dynamic light scattering, Raman spectroscopy, and X-ray diffractometry. In addition, the cytotoxicity of BP nanodots against C2C12 myoblasts was evaluated. Moreover, their cell imaging potential was investigated. Results: Herein, we concentrated on evaluating the cytotoxicity of BP nanodots and investigating their cell imaging potential. It was revealed that the BP nanodots were cytocompatible at a low concentration, although the cell viability was decreased with increasing BP nanodot concentration. Furthermore, our results demonstrated that the cells took up the BP nanodots, and the BP nanodots exhibited green fluorescence. Conclusions: In conclusion, our findings suggest that the BP nanodots have suitable biocompatibility, and are promising candidates as fluorescence probes for biomedical imaging applications.

A Study on the Mechanical Properties of Duplex Stainless Steel Weldment According to Mo Contents

  • Bae, Seong Han;Lim, Hee Dae;Jung, Won Jung;Gil, Woong;Jeon, Eon Chan;Lee, Sung Geun;Lee, Hyo Jong;Kim, In Soo;Lee, Hae Woo
    • 대한금속재료학회지
    • /
    • 제50권9호
    • /
    • pp.645-651
    • /
    • 2012
  • This study investigated changes in phase fraction caused by the addition of Mo, as well as the subsequent behaviour of N and its effect on the mechanical properties of welded 24Cr-N duplex stainless steel weld metals. Filler metal was produced by fixing the contents of Cr, Ni, N, and Mn while adjusting the Mo content to 1.4, 2.5, 3.5 wt%. The delta ferrite fraction increased as the Mo content increased. In contrast, the ${\gamma}$ fraction decreased and changed from a round to an acicular shape. Secondary austenite (${\gamma}^{\prime}$) was observed in all specimens in a refined form, but it decreased as the Mo content increased to the extent that it was nearly impossible to find any secondary austenite at 3.5 wt% Mo. Both tensile and yield strengths increased with the addition of Mo. In contrast, the highest value of ductility was observed at 1.41 wt% Mo. At all temperatures, impact energy absorption showed the lowest value at 3.5 wt% Mo, at which the amount of ${\delta}$-ferrite was greatest. There was no significant temperature dependence of the impact energy absorption values for any of the specimens. As the fraction of ${\gamma}$ phase decreased, the amount of N stacked in the ${\gamma}$ phase increased. Consequently, the stacking fault energy decreased, while the hardness of ${\gamma}$ increased.

Analysis of Impact on ERP Customization Module Using CSR Data

  • Yoo, Byung-Keun;Kim, Seung-Hee
    • Journal of Information Processing Systems
    • /
    • 제17권3호
    • /
    • pp.473-488
    • /
    • 2021
  • The enterprise resource planning (ERP) system is a standardized and advanced business process that many companies are implementing now-a-days through customization. However, it affects the efficiency of operations as these customizations are based on uniqueness. In this study, we analyzed the impact of customized modules and processing time on customer service request (CSR), by utilizing the stacked CSR data during the construction and operation of ERP, focusing on small and medium-sized enterprises (SMEs). As a result, a positive correlation was found between unit companies and the length of ERP implementation; ERP modules and the length of ERP implementation; ERP modules and unit companies; and the type of ERP implementation and ERP module. In terms of CSR, a comparison of CSR processing time of CBO (customized business object) module and STD (standard) module revealed that while the five modules did not display statistically significant differences, one module demonstrated a statistically very significant difference. In sum, the analysis indicates that the CBO-type CSR and its processing cost are higher than those of STD-type CSR. These results indicate that companies planning to implement an ERP system should consider the ERP module and their customization ratio and level. It not only gives the theoretical validity that should be considered as an indicator for decision making when ERP is constructed, but also its implications on the impact of processing time suggesting that the maintenance costs and project scheduling of ERP software must also be considered. This study is the first to present the degree of impact on the operation and maintenance of customized modules based on actual data and can provide a theoretical basis for applying SW change ratio in the cost estimation of ERP system maintenance.

Thoracodorsal artery flaps for breast reconstruction-the variants and its approach

  • Thomsen, Jorn Bo;Rindom, Mikkel Borsen;Rancati, Alberto;Angrigiani, Claudio
    • Archives of Plastic Surgery
    • /
    • 제48권1호
    • /
    • pp.15-25
    • /
    • 2021
  • Thoracodorsal artery (TDA) flaps, ranging from the vascular-pedicled thoracodorsal artery perforator (TDAP) flap, the propeller TDAP flap, and the muscle-sparing latissimus dorsi (MSLD) flap to the conventional latissimus dorsi (LD) flap and the extended LD flap, can all be used for breast reconstruction. The aim of this paper and review is to share our experiences and recommendations for procedure selection when applying TDA-based flaps for breast reconstruction. We describe the different surgical techniques and our thoughts and experience regarding indications and selection between these procedures for individual patients who opt for breast reconstruction. We have performed 574 TDA flaps in 491 patients: 60 extended LD flaps, 122 conventional LD flaps, two MSLD flaps, 233 propeller TDAP flaps, 122 TDAP flaps, and 35 free contralateral TDAP flaps for stacked TDAP breast reconstruction. All the TDA flaps are important flaps for reconstruction of the breast. The LD flap is still an option, although we prefer flaps without muscle when possible. The vascular-pedicled TDAP flap is an option for experienced surgeons, and the propeller TDAP flap can be used in most reconstructive cases of the breast, although a secondary procedure is often necessary for correction of the pedicle bulk. The extended LD flap is an option for women with a substantial body mass index, although it is associated with the highest morbidity of all the TDA flaps. The MSLD flap can be used if the perforators are small or if dissection of the perforators is assessed to be hazardous.

HPLC-MS/MS를 활용한 DTP(Digital Textile Printing) 제품의 지표성분 동시분석법 연구 - 천연염료를 중심으로 - (Study of Simultaneous Analysis of Indicator Components of DTP(Digital Textile Printing) Textile Products Using HPLC-MS/MS - Focusing on Natural Dyes -)

  • 이원경;성은지;문종렬;김미지;김종훈
    • 한국의류산업학회지
    • /
    • 제23권6호
    • /
    • pp.844-851
    • /
    • 2021
  • Due to the increase in consumers' interest about well-being, interest in eco-friendly products has been increasing due to the harmful effects of various harmful substances contained in textile products and environmental issues. As a result, natural dyes of less potential risk than synthetic dyes and digital textile printing(DTP) textile products with less environmental pollution are drawing attention. However, due to the lack of evaluation criteria for DTP textile products with natural ink and the nature of many colors are stacked layer by layer for dying, the need for simultaneous analysis is emerging. To evaluate whether the natural dye is derived from natural ingredients, the biocarbon content is analyzed. However, in the case of ink made using natural dyes and DTP textile products using natural ink, it is difficult to analyze the biocarbon content due to the limitation of the presence of a small amount of dye contained therein. In this study, we were shown the possibility of natural derived verification by cross-checking the analytes of natural dyes (Persicaria tinctoria, an indigo dye; Dactylopius coccus, a light red; and Curcum longa L., i.e., turmeric) and natural ink with HPLC-MS/MS. The coefficient of determination was 0.999 or higher, the limit of quantification was 0.647-3.664 ㎍/L and a %RSD of each indicator material was less than 10. Then, the extraction amount of natural dyes for five patterned fabrics was analyzed.

Machine Learning-based Classification of Hyperspectral Imagery

  • Haq, Mohd Anul;Rehman, Ziaur;Ahmed, Ahsan;Khan, Mohd Abdul Rahim
    • International Journal of Computer Science & Network Security
    • /
    • 제22권4호
    • /
    • pp.193-202
    • /
    • 2022
  • The classification of hyperspectral imagery (HSI) is essential in the surface of earth observation. Due to the continuous large number of bands, HSI data provide rich information about the object of study; however, it suffers from the curse of dimensionality. Dimensionality reduction is an essential aspect of Machine learning classification. The algorithms based on feature extraction can overcome the data dimensionality issue, thereby allowing the classifiers to utilize comprehensive models to reduce computational costs. This paper assesses and compares two HSI classification techniques. The first is based on the Joint Spatial-Spectral Stacked Autoencoder (JSSSA) method, the second is based on a shallow Artificial Neural Network (SNN), and the third is used the SVM model. The performance of the JSSSA technique is better than the SNN classification technique based on the overall accuracy and Kappa coefficient values. We observed that the JSSSA based method surpasses the SNN technique with an overall accuracy of 96.13% and Kappa coefficient value of 0.95. SNN also achieved a good accuracy of 92.40% and a Kappa coefficient value of 0.90, and SVM achieved an accuracy of 82.87%. The current study suggests that both JSSSA and SNN based techniques prove to be efficient methods for hyperspectral classification of snow features. This work classified the labeled/ground-truth datasets of snow in multiple classes. The labeled/ground-truth data can be valuable for applying deep neural networks such as CNN, hybrid CNN, RNN for glaciology, and snow-related hazard applications.

조적채움벽을 갖는 RC 벽-슬래브 골조의 내진성능 연구 (Investigation of Seismic Performance of RC Wall-Slab Frames with Masonry Infill)

  • 김찬호;이승제;허석재;엄태성
    • 한국지진공학회논문집
    • /
    • 제26권3호
    • /
    • pp.137-147
    • /
    • 2022
  • This study investigated the seismic performance of reinforced concrete (RC) wall-slab frames with masonry infills. Four RC wall-slab frames with or without masonry infill were tested under cyclic loading. The RC frames were composed of in-plane and out-of-plane walls and top and bottom slabs. For masonry infill walls, cement bricks were stacked applying mortar paste only at the bed joints, and, at the top, a gap of 50 mm was intentionally left between the masonry wall and top RC slab. Both sides of the masonry walls were finished by applying ordinary or fiber-reinforced mortars. The tests showed that despite the gap on top of the masonry walls, the strength and stiffness of the infilled frames were significantly increased and were different depending on the direction of loading and the finishing mortars. During repeated loading, the masonry walls underwent horizontal and diagonal cracking and corner crushing/spalling, showing a rocking mode inside the RC wall-slab frame. Interestingly, this rocking mode delayed loss of strength, and as a result, the ductility of the infilled frames increased to the same level as the bare frame. The interaction of masonry infill and adjacent RC walls, depending on the direction of loading, was further investigated based on test observations.

도심 항공 교통을 위한 전기동력 수직 이착륙기의 동축 동회전 로터의 성능해석 연구 (Study on Performance Analyses on Coaxial Co-rotating Rotors of e-VTOL Aircraft for Urban Air Mobility)

  • 이유빈;박재상
    • 한국항공우주학회지
    • /
    • 제49권12호
    • /
    • pp.1011-1018
    • /
    • 2021
  • 본 연구에서는 회전익기 통합해석 코드인 CAMRAD II를 이용하여 동축 동회전 로터(적층 로터)의 모델링과 제자리 비행 성능해석을 수행하였다. 본 연구는 주요 설계 변수인 방위각 위상차와 상/하단 로터 사이의 간격을 적절히 고려하여 동축 동회전 로터를 모델링하였으며, 기존 동축 반전 로터의 상단 로터의 토크 값을 동축 동회전 로터의 트림 목표 값으로 이용한 트림 기법 및 동축 반전 로터의 전체 추력 값을 트림 목표 값으로 이용한 트림 기법에서 제자리 비행 성능해석을 각각 수행하였다. 토크 트림 기법을 이용하였을 시 전체 로터의 추력은 -10°의 방위각 위상차에서 1.84% 최대화되었고, 추력 트림 기법을 적용하였을 때에는 전체 로터의 동력은 20°의 방위각 위상차에서 4.53%만큼 최소화되었다. 본 연구를 통하여 전기동력 수직 이착륙기에서 활용될 수 있는 동축 동회전 로터의 제자리 비행 성능이 방위각 위상차의 값에 따라 변화될 수 있음을 확인하였다.