检索结果-内蒙古大学图书馆

Asia and South Pacific design Automation Conference

作者： Yunhung Gao Hongyan Li Kevin Zhang Xueru Yu Hang Lu School of Electronics Engineering and Computer Science Peking University Beijing China State Key Lab of Computer Architecture Institute of Computing Technology CAS University of Chinese Academy of Sciences Beijing China Shanghai Integrated Circuits R&D Center Co. Ltd Shanghai China

Vanilla deep Neural Networks (dNN) after training are represented with native floating-point 32 (fp32) weights. We observe that the bit-level sparsity of these weights is very abundant in the mantissa and can be directly exploited to speed up model inference. In this paper, we propose Mortar, an off-line/on-line collaborated approach for fp32 dNN acceleration, which includes two parts: first, an off-line bit sparsification algorithm to construct the target formulation by “mantissa morphing”, which maintains higher model accuracy while increasing bit-level sparsity; second, the associating hardware accelerator architecture to speed up the on-line fp32 inference through manipulating the enlarged bit sparsity. We highlight the following results by evaluating various deep learning tasks, including image classification, object detection, video understanding, video & image super-resolution, etc.: We (1) increase bit-level sparsity up to 1.28~2.51x with only a negligible -0.09~0.23% accuracy loss, (2) maintain on average 3.55% higher model accuracy while increasing more bit-level sparsity than the baseline, (3)and our hardware accelerator outperforms up to 4.8x over the baseline, with an area of 0.031 mm 2 and power of 68.58mW.

关键词： deep learning Training Mortar Superresolution Software algorithms Inference algorithms Software

来源：评论

学校读者我要写书评

暂无评论

Multi-basis continuous wavelet transform feature wavelengths selection and machine learning with hyperspectral imaging for non-destructive prediction of 1,2-propylene glycol content in cased tobacco leaves

引用

Applied Optics 2025年第18期64卷 5087-5098页

作者： Zhongpan Yang Wuyin Jin Jinsong du School of Mechanical and Electronical Engineering Lanzhou University of Technology No. 287 Langongping Road Qilihe District Lanzhou City Gansu Province 730050 China Technology R&D Center Gansu Tobacco Industrial Co. Ltd. No. 1111 Nanbinhe Middle Road Qilihe District Lanzhou City Gansu Province 730050 China Key Laboratory of Tobacco Processing Technology of CNTC Zhengzhou Tobacco Research Institute of CNTC No. 2 Fengyang Street National High-Tech Industrial Development Zone Zhengzhou City Henan Province 450001 China

The process of manufacturing tobacco casings constitutes a critical step in the production of tobacco leaves, exerting a substantial influence on the improvement of their physical and chemical characteristics, and consequently, the quality of the final product. Nevertheless, the prevailing approach to casing accuracy detection is predominantly focused on dosage monitoring, with a paucity of attention being paid to real-time effect evaluation. Current hyperspectral-based detection systems encounter difficulties in extracting trace additive features and managing high-dimensional data under limited sample conditions. A multi-basis continuous wavelet transform (CWT) and machine-learning-integrated framework for non-destructive propylene glycol (PG) content prediction were proposed in this paper, addressing precision limitations in tobacco quality monitoring. The preprocessing of hyperspectral imaging data from six PG concentration levels were undertaken via Savitzky–Golay filtering, followed by multiscale decomposition using three CWT basis functions with morlet, Mexican hat, and Gaussian wavelets. A dual optimization mechanism combining correlation threshold filtering and wavelength frequency statistics was developed to enable efficient feature wavelength selection. Furthermore, a stacking regression model was constructed and compared with standalone algorithms. The results demonstrated that the multiscale combined MMG strategy achieved 79.86% dimensionality reduction by selecting 58 feature wavelengths covering adjacent regions in the near-infrared and short-wave infrared (NIr-SWIr) range, significantly enhancing model generalization compared to full-spectrum inputs. Additionally, the stacking regression model attained optimal performance, with a testing set coefficient of determination of 0.9704, under combined MMG input through synergistic complementarity of heterogeneous base learners, with a root mean square error of 0.3188. It is confirmed in this paper that spec

关键词： deep learning Hyperspectral imaging Machine learning Neural networks Spatial resolution Wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

A Versatile Task Allocation System for Agricultural Operations by Formulation with a Split delivery Vehicle routing Problem

A Versatile Task Allocation System for Agricultural Operatio...

引用

IEEE/ASME (AIM) International Conference on Advanced Intelligent Mechatronics

作者： Kenta Hayakawa Shunsuke Miyashita Nagahiro Fujiwara ryota Takamido Jun Ota Department of Precision Engineering School of Engineering The University of Tokyo 7-3-1 Hongo Bunkyo-Ku Tokyo Japan Technology Innovation R&D Dept.II KUBOTA Corporation 1-11 Takumi-cho Sakai-Ku Research & Development Head-quarters Sakai-shi Osaka Japan Center for Engineering (RACE) School of Engineering Research into Artifacts The University of Tokyo 7-3-1 Hongo Bunkyo-Ku Tokyo Japan

ISBN: (数字)9798350355369

ISBN: (纸本)9798350355376

In recent years, there has been a growing interest in smart agriculture, aimed at enhancing efficiency in farming through the use of technology. Farmers, working with multiple machines across various fields, necessitate the development of a versatile task assignment and route generation system (task allocation system) to cater to a diverse range of conditions. Previous studies have proposed various methods to address this problem. However, it remains unclear which method is most suitable when specific conditions tailored to individual farms are considered. Additionally, the performance of these methods is not fully utilized as hyperparameters are often set empirically. Consequently, the development of a truly versatile task allocation system for agriculture remains unachieved. Therefore, this study aims to develop a versatile task allocation system for agriculture. Specifically, we created “the field nodes divided graph” by dividing fields into the maximum number of vehicles operating in that field, and by formalizing it as a split delivery vehicle routing problem (SdVrP), we modeled the farmland considering the operation of multiple vehicles. Using computer simulations, this study identifies the most effective optimization methods and hyperparameter combinations under varying farm sizes and maximum calculation times. The results of computer simulations applying a local search method, a simulated annealing (SA), a genetic algorithm (GA), and an ant colony optimization (ACO) to five types of hypothetical grid-shaped farmlands revealed that the SA algorithm is superior for smaller farmlands, while the local search method excel in larger farmlands with shorter calculation times. Furthermore, the study derives the best hyperparameter values corresponding to different farm sizes.

关键词： Smart agriculture Search methods Computer simulation Metaheuristics Simulated annealing routing Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Influence Of Microstructural Assemblage Of Substrate On The Adhesion Strength Of Tisin Coated Pcbn Systems

Influence Of Microstructural Assemblage Of Substrate On The ...

引用

World PM 2022 Congress and Exhibition

作者： Gordon, Sandra Jiménez-Piqué, Emilio rodriguez-Suarez, Teresa M'Saoubi, rachid roa, Joan Josep Franca, Luis F.P. Llanes, Luis CIEFMA - Department of Materials Science and Engineering EEBE Universitat Politècnica de Catalunya Barcelona08019 Spain Barcelona Research Center in Multiscale Science and Engineering Universitat Politècnica de Catalunya BarcelonaTech Barcelona08019 Spain Ltd Global Innovation Centre Fermi Avenue OxfordshireOX11 0QR United Kingdom R&D Material and Technology Development Seco Tools AB Fagersta73782 Sweden

ISBN: (纸本)9781899072552

The coating adhesion strength of two TiSiN-coated polycrystalline cubic boron nitride (PcBN) grades is studied in this work. Microstructural assemblage of the PcBN substrate, described in terms of cBN content, binder chemical nature and average cBN grain size, is considered as the main experimental variable. In the present investigation a conical indentation test method is implemented to assess load levels at which cracking and delamination emerge, and their evolution with increasing load. Tests were conducted by using a conical diamond indenter of 200 µm in radius (rockwell C) and applying loads ranging from 98 to 980 N. They were complemented by an extensive and detailed SEM inspection of fracture micro-mechanisms (i.e. type of cracking pattern and delamination). results show that emergence and evolution of damage strongly depend on the substrate microstructural assemblage: the higher the cBN-content, the more resistant to both cracking and delamination is the coated system. © European Powder Metallurgy Association (EPMA)

关键词： Bond strength (materials)

来源：评论

学校读者我要写书评

暂无评论

An Approximate High Gain Observer for Speed-sensorless Estimation of Induction Motors

引用

IEEE/CAA Journal of Automatica Sinica 2019年第1期6卷 53-63页

作者： Yebin Wang Lei Zhou Scott A.Bortoff Akira Satake Shinichi Furutani IEEE Mitsubishi Electric Research Laboratories the Department of Mechanical Engineering Massachusetts Institute of Technology the Advanced Technology R&D Center Mitsubishi Electric Corporation

rotor speed estimation for induction motors is a key problem in speed-sensorless motor drives. This paper performs nonlinear high gain observer design based on the full-order model of the induction motor. Such an effort appears nontrivial due to the fact that the full-model at best admits locally a non-triangular observable form(NTOF), and its analytical representation in the NTOF can not be obtained. This paper proposes an approximate high gain estimation algorithm, which enjoys a constructive design, ease of tuning, and improved speed estimation and tracking performance. Experiments demonstrate the effectiveness of the proposed algorithm.

关键词： Induction motor industrial applications nonlinear state estimation speed-sensorless motor drive

来源：评论

学校读者我要写书评

暂无评论

Bifunctional membrane design based on thiol-functionalized triazine-containing covalent organic polymer for integrated mercury ion removal and monitoring

引用

Journal of Membrane Science 2025年 733卷

作者： du, Yongjian Shan, Houchao Jiang, Yunfei Zhang, Ziyi Jiao, Jinwen Wang, Mengsen Cai, di Zhang, rui Jiao, Tifeng State Key Laboratory of Metastable Materials Science and Technology Hebei Key Laboratory of Applied Chemistry Hebei Key Laboratory of Nanobiotechnology Hebei Key Laboratory of Heavy Metal Deep-Remediation in Water and Resource Reuse Yanshan University Qinhuangdao066004 China National Energy R&D Center for Biorefinery Beijing University of Chemical Technology No.15 North 3rd Ring East Road Beijing100029 China School of Chemistry and Chemical Engineering Hunan University of Science and Technology Xiangtan411201 China

This study showcases the versatility of covalent organic polymers (COPs) as a viable platform for the selective identification and efficient removal of mercury ions (Hg2+). A thiol-functionalized, triazine-based N-rich COP called TBN-S-SH was developed, featuring a synergistic integration of electron-deficient triazine and electron-rich thiol moieties. This unique structure endows TBN-S-SH with robust fluorescence characteristics, enabling high sensitivity and selectivity in detecting Hg2+ through fluorescence quenching. The adsorbent exhibits exceptional Hg2+ uptake capacity of 1107 mg g−1, demonstrating strong affinity and remarkably rapid kinetics, with an initial adsorption rate of 200 mg g−1 min−1. density functional theory analysis reveals that the efficacy of TBN-S-SH stems from its nucleophilic structure and stable configurations, facilitated by coordination and cation-π interactions between TBN-S-SH and Hg2+. Based on this, a tubular dual-layer bifunctional TBN-S-SH membrane was developed, exhibiting an excellent mercury ion retention of 99.7 % and a high water permeance of 397 L m−2 h−1 MPa−1. Notably, this separation system enables real-time monitoring of the concentration of mercury ions on the TBN-S-SH membrane grown on the inner wall of the ceramic tube, providing an indicator of potential mercury ion leakage. This study highlights the multifunctionality of COPs in environmental remediation. © 2025 Elsevier B.V.

关键词： Copolymers

来源：评论

学校读者我要写书评

暂无评论

Comparative Study on Adversarial Captcha with Three Target detection Models

Comparative Study on Adversarial Captcha with Three Target D...

引用

Chinese Automation Congress (CAC)

作者： Zhang Xin Wujun Xu Yu Suping Ling Jin College of Information Science and Technology Donghua University Shanghai College of Information Science and Technology Donghua University Engineering Research Center of Digitalized Textiles and Fashion Technology Ministry of Education Shanghai Intel Asia Pacific R&D Co. Ltd Shanghai

In order to distinguish whether the user is a human or a robot, prevent decryption of password, accounts and so on, three kinds of target detection model training for Adversarial Captcha are proposed. The process of training and evaluate-on of adversarial samples are described in detail in this paper. depth target detection model is selected as attack model, and Projected Gradient descent (PGd) algorithm is applied to generate adversarial samples based on three models: YOLOv4, Single Shot MultiBox detector (SSd) and Faster region-CNN (r-CNN). The samples are compared with clean samples under undirected attack. The results show that the captcha samples have successfully misled the neural network with high confidence, making them output incorrect results. The algorithm proposed in this paper can be used in many fields in the Internet to verify whether it is human or not.

关键词： Training Automation Neural networks Object detection Passwords detectors Internet

来源：评论

学校读者我要写书评

暂无评论

Ammonia as an alternative fuel for vehicular applications: Paving the way for adsorbed ammonia and direct ammonia fuel cells

引用

Journal of Cleaner Production 2022年第0期376卷

作者： Sonker, Muskan Tiwary, Saurabh Kr Shreyash, Nehil Bajpai, Sushant ray, Mainak Kar, Sanjay Kumar Balathanigaimani, M.S. Department of Chemical and Biomolecular Engineering Georgia Institute of Technology GA Atlanta30318 United States Department of Chemical and Biomolecular Engineering University of Houston HoustonTX United States Department of Chemical Biological and Materials Engineering University of Oklahoma NormanOK United States R&D and Innovation HPCL-Mittal Energy Limited Bathinda Punjab India Department of Management Sciences Rajiv Gandhi Institute of Petroleum Technology Jais India Department of Chemical Engineering and Biochemical Engineering Rajiv Gandhi Institute of Petroleum Technology Jais India

This article is a roadmap for researchers in this field to look for an efficient and reliable alternative for fossil fuels in the form of ammonia. We have discussed different storage techniques, with one of them being ‘solid ammonia storage method’ using MOFs and metal ammine complexes as a better option to deal with ammonia toxicity. Besides, we have discussed various approaches of direct and indirect ammonia being used in internal combustion engines as fuel or within fuel cells for creating a sustainable energy source for the onboard vehicular application. To discuss ammonia's feasibility as a commercial fuel, a detailed discussion on ammonia economy is important. So, in the later sections of the article, we have discussed ammonia economy, emerging technologies for ammonia production, cost analysis, investment, the economics of vehicle engines and fuel cells, and opportunities ahead. In the article, apart from just advocating the feasibility of using ammonia as a future fuel, the authors have also talked about its downsides, associated challenges, and how there are alternative pathways to use ammonia indirectly as well. The authors believe that in the literature, very few articles present such a broad discussion on ammonia's feasibility as a future fuel. Thus, we believe this article will help the readers in formulating their own opinion regarding ammonia and work on finding the right fuel for a sustainable future. © 2022

关键词： Fuel cells

来源：评论

学校读者我要写书评

暂无评论

Highly Concentrated Multivariate Zif-8 Mixed Matrix Hollow Fiber Membranes for Co2 Separation: Scalable Fabrication and Process Analysis

SSRN

引用

SSrN 2023年

作者： An, Heseong Jung, Wonho Shin, Ju Ho Shin, Min Chang Park, Jung Hoon Lee, Jinwon Lee, Jong Suk Department of Chemical and Biomolecular Engineering Sogang University 35 Baekbeom-ro Mapo-gu Seoul04107 Korea Republic of C1 gas refinery R&D center Sogang University 35 Baekbeom-ro Mapo-gu Seoul04107 Korea Republic of Department of Chemical & Biochemical Engineering Dongguk University 30 Pildong-ro 1gil Jung-gu Seoul04620 Korea Republic of

Here, we report on the fabrication of mixed matrix hollow fiber membranes (MMHFMs) for CO2 separation using highly-concentrated 6FdA-dAM/triple-ligand, amine-modulated zeolitic imidazolate framework-8 (TAZIF-8) (60/40 w/w) via a dip coating process. The application of a γ-alumina interlayer serves a dual purpose of reducing the infiltration of the 6FdA-dAM/TAZIF-8 coating solution into the porous α-alumina support, while also providing a smooth surface for the support. Importantly, the co-solvent comprising N-methyl-2-pyrrolidone and dichloromethane further inhibits the formation of defects by maintaining an appropriate balance of volatility and polarity. This balanced combination is crucial in promoting phase separation and facilitating the dispersion of nanofillers within the polymer matrix, enabling the formation of a thin defect-free selective layer. The subsequent 40 wt.% TAZIF-8 MMHFM attains superior CO2 separation performance to most existing MMHFMs (CO2 permeance of 648 GPU and CO2/N2 and CO2/CH4 selectivity of 39.0 and 34.2, respectively). Furthermore, it maintains excellent CO2 separation performance over 120 d owing to good interfacial interaction between 6FdA-dAM and TAZIF-8s. Last, but not least, our process analysis reveals that the introduction of the TAZIF-8 MMHFM into the first stage of a dual-stage membrane process can improve the total energy demand for post-combustion CO2 capture due to its excellent CO2/N2 selectivity. Our current study provides significant progress toward the industrial application of post-combustion CO2 capture. © 2023, The Authors. All rights reserved.

关键词： dip coating

来源：评论

学校读者我要写书评

暂无评论

An extended LuGre model for estimating nonlinear frictions in feed drive systems of machine tools

引用

Procedia CIrP 2022年 107卷 452-457页

作者： Tiandong Xi Tomoya Fujita Sebastian Kehne ryosuke Ikeda Marcel Fey Christian Brecher Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University Steinbachstr. 19 52074 Aachen Germany Mitsubishi Electric Corporation Advanced Technology R&D Center Tsukaguchi-honmachi 8-1-1 Amagasaki Hyogo 661-8661 Japan

Nonlinear frictions in feed drive axes affect the control accuracy and dynamic of the machine tools. Conventional dynamic friction models neglect the position-dependent friction, caused by manufacturing errors and misalignment as well as the environment. This paper proposes an extended LuGre friction model for estimating the dynamic and position-dependent friction effects. The conventional LuGre model is modified by a position-dependent term, which describes the nonlinear friction behavior of machine axes with long travel distance. An online data-driven identification approach is developed to fast parametrize and update the model. The suggested method is validated on a real machine tool.

关键词： friction feed drive system machine tool system identification

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：