检索结果-内蒙古大学图书馆

International symposium on Chinese Spoken Language Processing

作者： Jia-Jyu Su Chen-Yu Chiang Yue-Shan Chang Chao-Yin Lin Jiunn-Horng Kang Min-Yuh Day National Taipei University New Taipei City Taipei Medical University Taipei City

ISBN: (数字)9798331516826

ISBN: (纸本)9798331516833

this study explores personalized Mandarin ASR for the speech impaired. Sentence and short phrase tasks selected from the VoiceBank-2023 corpus are examined for amyotrophic lat-eral sclerosis and esophageal-speech speakers. Unadapted speaker-independent ASR model is trained with the NER-Pro corpus to serve as a pretrained model for the follow-up speaker-independent and dependent models fine-tuned with the CDSD and VoiceBank-2023 corpora. Several fine-tuning strategies are investigated. Experimental results show that the adapted speaker-dependent models attain average character error rates (CERs) ranging from 6.8% (normal) to 11.9% (mod-erate dysarthria) and 5.9% (normal) to 12.9% (moderate) for sentence and short phrase tasks, respectively. Last, we have implemented the fine-tuning strategies on a large corpus of a hereditary spastic paraplegia patient to evaluate frequently used sentence tasks under open-set and close-set conditions, resulting in CERs of 10.8% and 3.8%, respectively.

关键词： Training Adaptation models Target recognition Error analysis Buildings Training data Speech recognition Distance measurement data models Speech processing

来源：评论

学校读者我要写书评

暂无评论

Research on Cross-modal image-text Retrieval Method based on multi-module graph structure

Research on Cross-modal image-text Retrieval Method based on...

引用

IEEE International symposium on Information (IT) in Medicine and Education, ITME

作者： Fuling Wang Li Liu Hua Ji Xinchen Dai Shandong Normal University School of Information Science and Engineering Jinan China

ISBN: (数字)9798350356670

ISBN: (纸本)9798350356687

With the rapid development of artificial intelligence technology and the widespread application of big data, the amount of media data in the real world is showing an explosive growth trend, and a large number of heterogeneous data have emerged in various industries. Because the scale of data is continuously expanding, the semantic correspondence between images and text has become increasingly important. How to efficiently retrieve in the vast amount of image-text data has become a highly concerned issue. there are still two limitations in the existing cross-modal image-text retrieval methods: 1) Different modal data have different distributions and are uneven. 2) Semantics are complex and abstract, and there is still an insurmountable semantic gap between vision and language. In this paper, we have done the following research: A cross-modal retrieval method based on multi-module graph structure is proposed, including visual, word, and phrase graph structure modules. Due to the complexity of data and the interaction between images and text not only in semantic information but also in rich spatial location information, we propose a visual module that combines image features with location features to learn the relationship between semantics and space using GCN. the diversity of text descriptions makes it difficult to learn consistent text feature representations. For this reason, we first use GCN in the word graph structure module to capture the associations between words based on semantic dependency relationships. then, in the phrase graph structure module, GCN is used to infer the relationships between noun phrases. Finally, a gated fusion layer is used to filter and reduce redundant information interference. this method has performed well on the Mirflickr25K and NUS-WIDE datasets.

关键词： visualization Graph convolutional networks Semantics Process control Interference Logic gates Media Information filters Complexity theory Cross modal retrieval

来源：评论

学校读者我要写书评

暂无评论

Distributed Feature Extraction on Apache Spark for Human Action Recognition

Distributed Feature Extraction on Apache Spark for Human Act...

引用

International Conference on Ubiquitous Information Management and Communication (IMCOM)

作者： Nguyen Anh Tu thien Huynh-the Kok-Seng Wong Dinh-Mao Bui Young-Koo Lee Dept. Computer Science Nazarbayev University Nur-Sultan Republic of Kazakhstan ICT Convergence Research Center Kumoh National Institute of Technology Gumi South Korea Dept. Computer Science & Engineering Kyung Hee University Yongin South Korea

ISBN: (数字)9781728154534

ISBN: (纸本)9781728154541

Local feature extraction is one of the most important tasks to build robust video representation in human action recognition. Recent advances in computing visual features, especially deep-learned features, have achieved excellent performance on a variety of action datasets. However, the extraction process is computing-intensive and extremely time-consuming when conducting it on large-scale video data. Consequently, to extract video features over big data, most of the existing methods that run on single machine become inefficient due to the limit of computation power and memory capacity. In this paper, we propose the elastic solutions for feature extraction based on the Spark framework. Particularly, exploiting the in-memory computing capability of Spark, the process of computing features are parallelized by partitioning video data into videos or frames and place them into resilient distributed datasets (RDDs) for the subsequent processing. then, we present the parallel algorithms to extract the state-of-the-art deep-learned features on the Spark cluster. Subsequently, using the distributed encoding, the extracted features are aggregated into the global representation which is fed into the learned classifier to recognize actions in videos. Experimental results on a benchmark dataset demonstrate that our proposed methods can significantly speed up the extraction process and achieve the promising scalability performance.

关键词： Feature extraction Sparks Trajectory Task analysis Liquid crystal displays visualization Streaming media

来源：评论

学校读者我要写书评

暂无评论

Characterizing the convective heat transfer on stator ventilation ducts for large hydro generators with a neural network

引用

COMPEL-thE INTERNATIONAL JOURNAL FOR COMPUTATION AND MAthEMATICS IN ELECTRICAL AND ELECTRONIC ENGINEERING 2015年第5期34卷 1522-1536页

作者： Schrittwieser, M. Biro, O. Farnleitner, E. Kastner, G. Graz Univ Technol Inst Fundamentals & Theory Elect Engn A-8010 Graz Austria Christian Doppler Lab Multiphys Simulat Anal & De Graz Austria Andritz Hydro GmbH R&D Dept Weiz Austria

Purpose - the purpose of this paper is to approximate the convective heat transfer using a few non-dimensional parameters in the design process of large synchronous machines. the computed convective wall heat transfer coefficient can be used in circuit models or can be defined in numerical heat conduction (HC) models to compute the thermal field in the solid domains without the time consuming computation of the fluid domain. Design/methodology/approach - Computational fluid dynamics (CFD) has been used to include a wide range of different designs, operating conditions and cooling schemes to ensure accurate results for a wide range of possible machines. Neural networks are used to correlate the computed heat transfer coefficients to various design parameters. the data set needed to define the weights and bias layers in the network has been obtained by several CFD simulations. A comparison of the evaluated solid temperatures with those obtained using the conjugate heat transfer (CHT) method has been carried out. the results have also been validated with calorimetric measurements. Findings - the validation of the HC model has shown that this model is capable of yielding accurate results in a few minutes, in contrast to the several hours needed by the CHT solution. the workflow to determine the convective heat transfer in a specific part of an electrical machine has been also been established. the approximation of the convective wall heat transfer coefficient is shown to be obtainable in sufficient detail by using a neural network. Research limitations/implications - the paper describes a method to approximate the convective heat transfer accurately in a few seconds, which is very useful in the design process. the heat convection can then be characterized in a HC model including the solid domains only. the losses can be defined as sources in the solid domains, e.g. copper and iron, obtained by electromagnetic calculations and the thermal field can hence be easily comput

关键词： Correlation analysis Neural networks Electrical machines thermal management thermal analysis Cooling systems Fluid dynamics

来源：评论

学校读者我要写书评

暂无评论

Unveiling the Conceptual analysis in the Direction of Developing a Fusion Virtual Reality Model for Hearing-Impaired Visual Interaction

Unveiling the Conceptual Analysis in the Direction of Develo...

引用

International Conference on Computer Applications and Industrial Electronics (ICCAIE)

作者： Ermiera Shafika Mokhtar Nurulnadwan Aziz Siti Salbiah Hamzah Siti Zulaiha Ahmad Ahmad Affandi Supli Azizah Che Omar College of Computing Informatics and Mathematics Universiti Teknologi MARA Cawangan Terengganu Dungun Terengganu Malaysia Faculty of Business and Management Universiti Teknologi MARA Cawangan Terengganu Dungun Terengganu Malaysia College of Computing Informatics and Mathematics Universiti Teknologi MARA Cawangan Perlis Arau Perlis Malaysia School of Computing and Data Science Xiamen University Malaysia Sepang Selangor Malysia School of Multimedia Technology and Communication Universiti Utara Malaysia Sintok Kedah Malysia

ISBN: (数字)9798350348798

ISBN: (纸本)9798350348804

Virtual Reality (VR) stands out as a rapidly evolving wearable technology. However, findings from previous studies reveal that existing VR models exhibit a significant deficiency in addressing the visual interaction requirements of the hearing-impaired (HI) population. While some studies focus on developing VR for HI individuals and incorporating assistive technology (AT) concepts, however, the integration of AT, empathic design (ED), and visual interaction (VI) principles remain underexplored. this study addresses the gap by proposing a fusion VR model that prioritizes the needs of HI individuals. the challenges faced by the HI, such as difficulties in visually utilizing sign language, gestures, body language, and lip sync, underscore the importance of adopting AT, ED, and VI design principles. through a rapid review approach, this paper strengthens the research foundation and outlines a conceptual framework for the development of a fusion VR model. this study aligns with the 12th Malaysian Plan's Pillar 5, emphasizing the empowerment of specific target groups, and supports SDG 10 by promoting social equality for the HI. the proposed fusion VR model aims to enhance existing designs by integrating AT, ED, and VI principles, providing a valuable reference for researchers in the field.

关键词： Solid modeling visualization Analytical models Systematics Sociology Virtual reality Auditory system

来源：评论

学校读者我要写书评

暂无评论

Efficient Finite-Difference Time-Domain Method for Solving Electromagnetic Scattering in Layered Anisotropies

Efficient Finite-Difference Time-Domain Method for Solving E...

引用

International symposium on Antennas, Propagation and EM theory (ISAPE)

作者： Yilin Kang Yuxian Zhang Xiaoli Feng Lixia Yang Zhixiang Huang Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education and with the Information Materials and Intelligent Sensing Laboratory of Anhui Province Industry-Education-Research Institute of Advanced Materials and Technology for Integrated Circuits and with School of Electronic and Information Engineering Anhui University Hefei China

ISBN: (数字)9798350353129

ISBN: (纸本)9798350353136

In this article, we propose an efficient numerical method for layered models based on the conventional finite difference time-domain (FDTD) method. the three-dimensional (3-D) layered model is compressed into one-dimensional (1-D) calculation, which improves the problem of large space memory and long running time for 3-D modeling while ensuring the accuracy and feasibility of the algorithm. On the one hand, by employing the transformation relationship between the transverse wave vector and the field in space, the expanded Maxwell's rotation equations are simplified, so that the equation system only retains the spatial transformation in the propagation direction. this processing allows for the calculation of the propagation characteristics of layered media without requiring more computational resources. On the other hand, to ensure the universality of the algorithm, a fully tensor anisotropic medium is selected for detecting the propagation data. At the same time, we design the layered structure and conduct the consistent 3-D modeling in the simulation software COMSOL. the results showed that our proposed efficient numerical algorithm greatly reduced the loss of computing resources while ensuring accuracy.

关键词： Solid modeling three-dimensional displays Accuracy Tensors Software algorithms Vectors Mathematical models Numerical models Time-domain analysis Finite difference methods

来源：评论

学校读者我要写书评

暂无评论

Reorganization of gene network of Bacillus subtilis for amylase production and secretion under different conditions

Reorganization of gene network of Bacillus subtilis for amyl...

引用

the 14th China-Japan-Korea Joint symposium on Enzyme Engineering

作者： Shaomin Yan Guang Wu State Key Laboratory of Non-food Biomass Enzyme Technology National Engineering Research Center for Non-food BiorefineryGuangxi Biomass Industrialization Engineering InstituteGuangxi Key Laboratory of BiorefineryGuangxi Academy of Sciences

As an important cell factory,*** can produce large amounts of designed proteins,for which many efforts have been made to improve its properties in order to meet the demand for industrial *** has a sophisticated genetic system to regulate and control its protein production and *** ***,about 150 genes have been identified in response to amylase overproduction and super-secretion,meanwhile these genes can function differently under different conditions because no single gene acts alone without interaction with other *** it is necessary to study how these genes behave under different conditions in order to better understand this sophisticated genetic *** way to study how genes behave under different conditions is to conduct the transcriptome studies,whose data have been documented in Gene Expression Omnibus(GEO).In this study,we apply network analysis to investigating the transcriptome data of *** from GEO including platforms GPL 188,GPL343,GPL6257,GPL 15025,GPL 19890,*** different conditions,and construct gene networks focusing on the abovementioned genes related to amylase production and *** results show that gene networks are different under different experimental conditions in terms of gene clusters,which consequently lead the genes related to amylase production and secretion to change their *** changes do not act alone but are associated with different genes,thus the network analysis reveals what types of genes are more likely to follow amylase production and secretion genes to change their ***,gene hierarchical structures are built,and relevant implications are *** our knowledge,this is the first study to use network analysis to integrate all *** genes together to get a whole picture of the gene relationships for amylase production and secretion,which provides information on better understanding of the mechanisms as suggested by Big

关键词： Amylase Bacillus subtilis Gene Network analysis Overproduction Reorganization Hyper-secretion

来源：评论

学校读者我要写书评

暂无评论

Layer-Adaptive Low-Rank Adaptation of large ASR Model for Low-Resource Multilingual Scenarios

Layer-Adaptive Low-Rank Adaptation of Large ASR Model for Lo...

引用

International symposium on Chinese Spoken Language Processing

作者： Yi Han Hang Chen Jun Du Chang-Qing Kong Shi-fu Xiong Jia Pan University of Science and Technology of China Hefei iFlytek Research Hefei

ISBN: (数字)9798331516826

ISBN: (纸本)9798331516833

Fine-tuning pre-trained ASR models is a practical approach for multilingual scenarios. Especially when facing the scarcity of annotated data, Low-Rank Adaptation (LoRA) exhibits commendable efficiency in this regard. However, when the available resources are further reduced, LoRA algorithms often suffer from severe overfitting due to the need to fine-tune all parameters, thereby compromising performance. In response to this challenge, we propose enhancing the LoRA algorithm by fine-tuning only a subset of parameters. Specifically, we introduce three manual layer selection strategies to the LoRA algorithm: Attention-Wise LoRA (AW-LoRA), Value-Output LoRA (VO-LoRA), and Rank-1 LoRA (R1-LoRA). Furthermore, we develop a Layer-Adaptive LoRA (LA-LoRA) that automatically assesses and selects critical layers based on the gradient's second moments. Experimental results on 4-hour datasets validate that our VO-LoRA achieves a comparable Word Error Rate (WER) to the raw LoRA algorithm while requiring only 27.25% parameters to be fine-tuned. the LA-LoRA algorithm further reduces the number of parameters needing fine-tuning to 10% and achieves a relative WER reduction of 1.82%. Moreover, Japanese and Korean experiments demonstrate that AW-LoRA, VO-LoRA, R1-LoRA, and LA-LoRA algorithms exhibit strong generalization capabilities across different languages. Our code is available at https://***/proudpie/LA-LoRA.

关键词： Adaptation models Codes Error analysis Manuals Multilingual Matrix decomposition Overfitting

来源：评论

学校读者我要写书评

暂无评论

Low-Cost Commercial Integrated Spectral Sensor for Revealing Breast Cancer Margins

Low-Cost Commercial Integrated Spectral Sensor for Revealing...

引用

Electrical Engineering (ICEENG), 12th International Conference on

作者： Mohamed Hisham Fouad Aref Ibrahim H. Aboughaleb Mohamed A. Abbass Abdallah Abdelkader Hussein Sara Abd El-Ghaffar Yasser M. Sabry Biomedical Engineering Researcher Egyptian Armed Forces Cairo Egypt Biomedical Engineering Dept. Military Technical College Egypt Histopathology Dept. Kobri El-Koba Medical Complex Cairo Egypt Radiology Dept. Maadi Armed Forces Medical Complex Cairo Egypt Faculty of Engineering Ain Shams University Cairo Egypt

ISBN: (数字)9798350343427

ISBN: (纸本)9798350343434

Breast malignancy (BM) poses a significant global health challenge, impacting millions worldwide. Precise detection of BM margins is essential in clinical pathology assessments, aiding in surgical margin delineation to effectively reduce recurrence risks. Our proposal involves implementing a cost-effective commercial optical imaging system tailored for distinguishing between malignant and normal tissues using infrared (IR) wavelengths. this system incorporates a low-cost, small and non-invasive spectral sensor (NeoSpectra-Micro) based on Fourier transform IR principles utilizing monolithic MEMS technology, operating within the 1350 to 2500 nm range. Employing this setup, we captured spectral signatures from 30 malignant and 30 normal tissue samples sourced from diverse patient cohorts, analyzing them with detector software. Subsequently, we normalized the data by subtracting the mean value to eliminate offset and scaled it to a range of (-1 to 1) by dividing by the maximum value. these steps enable clear differentiation between normal and BM tissues based solely on pattern, minimizing amplitude effects. Diffused reflected spectra exhibited notable visual differences between BM and normal tissues across three ranges: 1580–1630 nm, 1720–1740 nm, and 2150–2205 nm. through data analysis, we identified optimal wavelength ranges of 1588 nm, 1722 nm, and 2195 nm, representing maximum differences between normal and tumor tissues. this study underscores the IR-optical sensor's potential as an effective adjunct to clinical pathology. Nonetheless, integration into current protocols necessitates large-scale clinical studies with statistically significant sample sizes.

关键词： Micromechanical devices Pathology visualization Protocols Fourier transforms Surgery Optical imaging

来源：评论

学校读者我要写书评

暂无评论

NoC Symbiosis (Special Session Paper)

NoC Symbiosis (Special Session Paper)

引用

International symposium on Networks-on-Chip (NOCS)

作者： Daniel Petrisko Chun Zhao Scott Davidson Paul Gao Dustin Richmond Michael Bedford Taylor Bespoke Silicon Group University of Washington

ISBN: (数字)9781728188478

ISBN: (纸本)9781728188485

Conventional wisdom states that Network-on-Chip router area grows quadratically with the channel width, and this perception has fundamentally shaped the assumptions of thousands of NoC papers that have been written to date, and many chip designs. However, this assumption is not entirely true. Simple analysis and empirical data from this paper shows that, in modern standard cell technology, a router's standard cell logic area actually grows only linearly; it is solely the wire routing area that grows *** we think of a NoC as a standalone block as is done in standard hierarchical VLSI design, then the overall area growth is indeed quadratic. But this approach either vastly under-utilizes logic area, or, in designs that match wire and logic area, leads to small network links. At the same time, many standard non-NoC logic blocks like processors or accelerator blocks typically use the standard cell logic area but need only a fraction of available wiring *** propose an alternative approach, NoC Symbiosis, in which router logic and the node logic it services are jointly placed together. the router absorbs excess wiring resources from the node logic, and the node logic absorbs excess standard cell area from the router. Current-day automatic place and route (APR) tools already automatically distribute the router logic across the node logic, in order to provide enough space for the wiring resources. With this approach, future SoC's can leverage vastly larger amounts of wiring bandwidth than ever before, or alternatively, reduce the area overhead of existing *** describe how we first encountered this phenomena, perform experiments to demonstrate its behavior, and provide design tips to help teams realize the potential of NoC Symbiosis.

关键词： Symbiosis Wiring Wires Network-on-chip Very large scale integration Tools Standards

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：