检索结果-内蒙古大学图书馆

IEEE International Symposium on Circuits and systems

作者： Chao Fang Liulu He Haonan Wang Jinghe Wei Zhongfeng Wang School of Electronic Science and Engineering Nanjing University Nanjing China School of Electrical and Computer Engineering University of Southern California Los Angeles CA USA China Key System & Integrated Circuit Co. Ltd China

Three-dimensional convolutional neural networks (3D CNNs) have attracted great attention in many complex computer vision tasks. However, it is difficult to deploy 3D CNNs on practical applications due to high algorithmic complexity, imposing the urgent requirement for dedicated accelerators. In this paper, F3D, a fast algorithm for 3D CNNs, is proposed based on 3D Fast Fourier Transform (FFT) and achieves a significant algorithmic strength reduction. We then propose an F3D-based hardware architecture, featuring a flexible FFT module and an efficient partial sum aggregation module. Furthermore, a dataflow for efficient mapping of 3D CNNs is designed, leading to a significant reduction of memory access. To demonstrate the efficiency of the above-mentioned techniques, we implement the widely used 3D CNN model, C3D, as our benchmark on the Xilinx VC709 platform. The experimental result shows that compared with the state-of-the-art accelerator, our work achieves a considerable throughput up to 864.1 GOPs, along with 1.68x and 2.00 x efficiency improvement on energy and DSP utilization, respectively.

关键词： Three-dimensional displays Fast Fourier transforms Throughput Hardware Convolutional neural networks Acceleration Task analysis

来源：评论

学校读者我要写书评

暂无评论

Students Online Exam Proctoring: A Case Study Using 360 Degree Security Cameras

Students Online Exam Proctoring: A Case Study Using 360 Degr...

引用

2020 International Conference on Emerging Technology in Computing, Communication and Electronics, ETCCE 2020

作者： Turani, Aiman A Alkhateeb, Jawad H Alsewari, AbdulRahman A. Taibah University Information System Department AlMadinah Almunawarah Saudi Arabia Taibah University Computer Science Department AlMadinah Almunawarah Saudi Arabia Faculty of Computing Software Engineering Department Pekan Pahang Malaysia

ISBN: (纸本)9781665419628

Online courses, online exams and online certificates are conducted by various universities and Information Technology (IT) institutes worldwide. Delivery tools have been created for conducting the exams from any place. Applying this will lead saving time and travelling cost. Nowadays, due to the COVID-19 pandemic, there is a big demand on the online courses and exams. This paper introduces a new approach for exam proctoring using 360-degree security camera. Mainly, online exams' security is a major concern. Thus, a delivery tools must not only ensure the identity of a test-taker but also the overall test integrity. In this paper, the usage of the 360-degree security camera over the traditional webcam was investigated in order to enhance the exam security and to minimize the stressful restrictions. To verify this goal, a case study on a group of volunteer students within the college of computer science and engineering was made. In addition, an automated proctoring model that will eliminate the need for a real-time proctoring and remove any scheduling constraints in order to prevent cheating is proposed in this paper. The machine learning algorithms is exploited to enrich the proposed system. A secure frame work using the biometric is applied in order to ensure authentication and running the online exam smoothly. © 2020 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

A Narrow Linewidth Terahertz Dual-Comb Spectrometer with an All-Fiber Adaptive Clock and a Real-Time Feedback Loop

SSRN

引用

SSRN 2022年

作者： Xia, Yu Li, Min Liu, Zheng Liu, Dikai Bai, Songtao He, Mingyang Shen, Xuling Yang, Kangwen Yuan, Shuai Yan, Ming Huang, Kun Zeng, Heping Shanghai Key Laboratory of Modern Optical System Engineering Research Center of Optical Instrument and System Ministry of Education School of Optical-Electrical and Computer Engineering University of Shanghai for Science and Technology Shanghai200093 China Jinan Institute of Quantum Technology Shandong Jinan250101 China State Key Laboratory of Precision Spectroscopy East China Normal University Shanghai200062 China Chongqing Key Laboratory of Precision Optics Chongqing Institute of East China Normal University Chongqing401121 China

Terahertz comb spectroscopy offers advantages of high-resolution in broadband terahertz regions. However, conventional implementations of terahertz comb spectrometers are difficult to simultaneously achieve high precision and long-term stability due to the residual timing jitter of the involved light sources. In this work, an adaptive-clock signal from optical beat note in optical region was extracted for real-time correcting the unstable sampling interval-time caused by the residual timing jitter and a loosely active frequency-stabilized scheme functioned as a real-time feedback. And finally, a long-term stable terahertz dual-comb spectrometer was constructed with a spectral resolution of 66.14 MHz, a comb linewidth of 1.9 kHz, a fractional precision of 1.7×10−9 and a long-term operation over 1200 hours. This proved that the precise acquisition of terahertz pulses could be realized through the compensation of timing jitter for pumped optical pulses. The presented terahertz comb spectrometer is featured with enhanced robustness compared to previous schemes based on two fully stabilized mode-locked laser combs, which would thus facilitate field deployment outside the laboratories. © 2022, The Authors. All rights reserved.

关键词： Spectrometers

来源：评论

学校读者我要写书评

暂无评论

MergeNAS: Merge operations into one for differentiable architecture search 29

MergeNAS: Merge operations into one for differentiable archi...

引用

29th International Joint Conference on Artificial Intelligence, IJCAI 2020

作者： Wang, Xiaoxing Xue, Chao Yan, Junchi Yang, Xiaokang Hu, Yonggang Sun, Kewei MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University China Department of Computer Science and Engineering Shanghai Jiao Tong University China IBM Research - China IBM System

ISBN: (纸本)9780999241165

Differentiable architecture search (DARTS) has been a promising one-shot architecture search approach for its mathematical formulation and competitive results. However, besides its caused high memory utilization and a large computation requirement, many research works have shown that DARTS also often suffers notable over-fitting and thus does not work robustly for some new tasks. In this paper, we propose a one-shot neural architecture search method referred to as MergeNAS by merging different types of operations e.g. convolutions into one operation. This merge-based approach not only reduces the search cost (about half a GPU day), but also alleviates over-fitting by reducing the redundant parameters. Extensive experiments on different search space and various datasets have been conducted to verify our approach, showing that MergeNAS can converge to a stable architecture and achieve better performance with fewer parameters and search cost. For test accuracy and its stability, MergeNAS outperforms all NAS baseline methods implemented on NASBench-201, including DARTS, ENAS, RS, BOHB, GDAS and hand-crafted ResNet. © 2020 Inst. Sci. inf., Univ. Defence in Belgrade. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Photodynamic effect of indocyanine green and its application to hepatocellular carcinoma treatment

引用

Journal of Cancer Metastasis and Treatment 2024年第1期10卷 169-185页

作者： Ester Jo Kush Savsani Anjelica Alfonso Andrew Park Seung Duk Lee Devanand Sakar Nathaniel Kinsey Yuzuru Sambommatsu Daisuke Imaid Aamir Khan Amit Sharma Muhammad Saeed Vinay Kumaran Adrian Cotterell Marlon Levy David Bruno College of Humanities and Sciences Virginia Commonwealth UniversityRichmondVA 23220USA Department of Surgery Virginia Commonwealth UniversityRichmondVA 23298USA School of Medicine Virginia Commonwealth UniversityRichmondVA 23298USA Department of Transplant Surgery Virginia Commonwealth University Health SystemRichmondVA 23219USA Department of Human and Molecular Genetics Virginia Commonwealth UniversityRichmondVA 23298USA Department of Electrical and Computer Engineering Virginia Commonwealth UniversityRichmondVA 23284USA

Hepatocellular carcinoma(HCC)stands as the primary cause of liver cancers,with limited treatment options outside of surgical resection or *** therapy(PDT)using indocyanine green(ICG)as a photosensitizer offers a promising therapeutic option for *** PDT has demonstrated efficacy in vitro and in vivo,providing a safe treatment option for patients ineligible for resection or *** in PDT optimization include limited light delivery,poor photosensitizer optimization,and low oxygen *** solutions,such as ICG incorporation into nanostructures and direct administration for HCC cases,show *** ICG PDT with Lentinula edodes mycelia(LEM)and hydrogen gas inhalation may address oxygen production *** a combination of therapeutic approaches and therapies may increase the efficacy of ICG PDT,making it the safer and more effective treatment option for ***’s proven clinical safety may expedite potential approval for its use in *** in photosensitizer technology,such as aggregation-induced emission photosensitizers(AIE-PSs),and combination therapies will enhance PDT efficacy shortly,as developments show promise for improved *** HCC patients,ICG PDT presents a valuable therapeutic *** work is novel because it explores new combination therapy approaches,leveraging the benefits of PDT with ICG and the developments in photosensitizer technology,nanotechnology,and the combination of other treatments such as LEM and hydrogen gas,which could potentially revolutionize the treatment of HCC and offer a better prognosis for patients.

关键词： Biomarkers cancer therapy hepatocellular carcinoma photodynamic therapy indocyanine green ICG HCC PDT

来源：评论

学校读者我要写书评

暂无评论

Scaling Limitations of Line TFETs at Sub-8-nm Technology Node

Scaling Limitations of Line TFETs at Sub-8-nm Technology Nod...

引用

2020 International Symposium on VLSI Technology, systems and Applications, VLSI-TSA 2020

作者： Thoti, Narasimhulu Li, Yiming Kola, Sekhar Reddy Parallel and Scientific Computing Laboratory Eecs International Graduate Program Institute of Communications Engineering Department of Electrical and Computer Engineering Center for MmWave Smart Radar System and Technologies National Chiao Tung University 1001 Ta-Hsueh Rd Hsinchu300 Taiwan

ISBN: (纸本)9781728142326

The scope of the work is to investigate limitations in device scaling by identifying various parameters of short channel effects (SCEs) in current challenging geometries of the line tunnel field effect transistors (TFETs). Key factors of the considered device, such as the doping and the thickness (tn) of n-epitaxial region, and source-to-drain length (LSDeff) scaling cannot be tuned anymore to boost the device characteristics for the sub-8-nm technology node. The main results of this study indicates that the engineering acceptable performance are achieved at a low doping of 5 1018 cm-3, an optimal tn as low (about 0.5 nm), and a LSDeff greater than 12.5 nm. Hence, the line TFETs below sub-8-nm faces serious bottleneck of scaling and cannot be further scaled with the conventional scaling rule at all. © 2020 IEEE.

关键词： Tunnel field effect transistors

来源：评论

学校读者我要写书评

暂无评论

Hard Sample Matters a Lot in Zero-Shot Quantization

Hard Sample Matters a Lot in Zero-Shot Quantization

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Huantong Li Xiangmiao Wu Fanbing Lv Daihai Liao Thomas H. Li Yonggang Zhang Bo Han Mingkui Tan South China University of Technology Information Technology R&D Innovation Center of Peking University Changsha Hisense Intelligent System Research Institute Co. Ltd School of Electronic and Computer Engineering Peking University Shenzhen Graduate School Shenzhen China Hong Kong Baptist University Key Laboratory of Big Data and Intelligent Robot Ministry of Education PengCheng Laboratory

Zero-shot quantization (ZSQ) is promising for compressing and accelerating deep neural networks when the data for training full-precision models are inaccessible. In ZSQ, network quantization is performed using synthetic samples, thus, the performance of quantized models depends heavily on the quality of synthetic samples. Nonetheless, we find that the synthetic samples constructed in existing ZSQ methods can be easily fitted by models. Accordingly, quantized models obtained by these methods suffer from significant performance degradation on hard samples. To address this issue, we propose HArd sample Synthesizing and Training (HAST). Specifically, HAST pays more attention to hard samples when synthesizing samples and makes synthetic samples hard to fit when training quantized models. HAST aligns features extracted by full-precision and quantized models to ensure the similarity between features extracted by these two models. Extensive experiments show that HAST significantly outperforms existing ZSQ methods, achieving performance comparable to models that are quantized with real data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

The Approach of Establishing the High-Resolution Shading Occupant Behavior Model in the Office Building

SSRN

引用

SSRN 2023年

作者： Chen, Gaoxiang Lu, Jun Zhang, Zhiang Deng, Wu Bie, Jing Yao, Jian Yan, Lei Guan, Rui Xia, Bowei Zhang, Jialu Yang, Rouyao Department of Architecture and Built Environment University of Nottingham Ningbo China Ningbo315100 China Department of Architecture Ningbo University Ningbo China Department of Financial Engineering Ningbo University Ningbo China Department of Information Management and Information System Ningbo University Ningbo China Department of Computer Science University of Nottingham Ningbo China Ningbo315100 China Ningbo University China

The environmental behavior of building occupants has become the key factor in determining indoor thermal comfort and building energy performance. The diversity and randomness of occupant behavior may lead to variations in building energy consumption. Early studies on shading occupant behavior have increasingly involved mathematical tools to establish a mathematical model for correlating shading behavior patterns and energy consumption. However, the existing model does not consider the multiple states of shading (only fully open, fully closed, or half-open state, which has a large deviation from the actual situation), accessibility of shading behavior (ASB), room configuration (single people room and multi-people room), occupancy rate and other factors, resulting in low accuracy of the model. This study proposed an index of the ASB to analyze and explore the characteristics of blind occupant adjustment behavior (named shading behavior afterwards). It also promotes an occupant-centric consideration in occupant behavior studying. By using the accessibility classification of shading behavior with building parameters, blinds can be quickly divided into four types: high ASB, medium ASB, general ASB, and low ASB. Based on the results of shading classification and integrating the data mining techniques such as data feature selection, soft voting, and Markov method was integrated, a high-resolution shading occupant behavior (HRSOB) model is proposed. Using the measured data of office buildings on a case study campus, the model selects the optimal algorithm and evaluates and verifies the HRSOB model. The results show that the predictive performance of the HRSOB model is improved by 10.92% compared with the typical SOB model. Finally, the advantages and disadvantages of the model are discussed. © 2023, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

(Keynote) Terahertz Nanoplasmonics Technology: Physics, Applications, and Commercialization 237

(Keynote) Terahertz Nanoplasmonics Technology: Physics, Appl...

引用

237th ECS Meeting with the 18th International Meeting on Chemical Sensors, IMCS 2020

作者： Shur, Michael Department of Electric Computer and System Engineering Department of Physics Applied Physics and Astornomy Rensselaer Polytechnic Insititute TroyNY12180 United States Electronics of the Future Inc VienneVA22181 United States

ISBN: (纸本)9781607688952

Nanoscale Field Effect Transistors (FETs) operating in the Terahertz (THz) range of frequencies rely on the excitation and rectification of the oscillations of the electron density in the device channels - plasma waves. TeraFETs are expected to enable numerous potential applications in THz sensing, imaging, and communications. A new emerging application is for cyber hardware security to distinguish between faked and genuine VLSI chips and predict the VLSI reliability. © 2020 ECS - The Electrochemical Society.

关键词： Field effect transistors

来源：评论

学校读者我要写书评

暂无评论

Multi-constraint UAV Fast Path Planning Based on Improved A∗ Algorithm 2

Multi-constraint UAV Fast Path Planning Based on Improved A...

引用

2020 2nd International Conference on computer Modeling, Simulation and Algorithm, CMSA 2020

作者： Kong, Xiaodi Pan, Bin Cherkashin, Evgeny Zhang, Xiaoyang Liu, Linke Hou, Jian School of Computer and Communication Engineering Liaoning ShiHua University Fushun China School of Science Liaoning ShiHua University Fushun China Matrosov Institute for System Dynamics Control Theoiry of SB RAS Novosibirsk Russia

Path planning is an important part of UAV intelligent control technology. For the current UAV path planning, the A∗ algorithm has a large memory overhead during the planning process, and the search speed is slow. It cannot meet the path planning in a complex three-dimensional environment. The immediacy requirement and considering the constraints are less, this paper made the following improvements to the A∗ algorithm. The first, combining an anytime repair search framework with a weighted A∗ algorithm, it is possible to quickly find a feasible path during the search process. The second, aiming at the defect of relying on low heuristic function weights in the algorithm, a double ranking criterion is proposed to improve the efficiency of the algorithm approaching the optimal path in the iterative process. And third, to reduce the number of node expansions in the planning process, increase list storage constraints have been improved. Finally, the results of simulation experiments show that the improved algorithm proposed in this paper can quickly generate feasible paths, and can continuously optimize the approach to the optimal path within a specified time, which is far superior to the traditional A∗ algorithm in terms of planning efficiency and immediacy. © Published under licence by IOP Publishing Ltd.

关键词： Motion planning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：