检索结果-内蒙古大学图书馆

Vector Symbolic Architectures as a computing framework for Emerging Hardware

PROCEEDINGS OF THE IEEE 2022年第10期110卷 1538-1571页

作者： Kleyko, Denis Davies, Mike Frady, E. Paxon Kanerva, Pentti Kent, Spencer J. Olshausen, Bruno A. Osipov, Evgeny Rabaey, Jan M. Rachkovskij, Dmitri A. Rahimi, Abbas Sommer, Friedrich T. Univ Calif Berkeley Redwood Ctr Theoret Neurosci Berkeley CA 94720 USA Res Inst Sweden Intelligent Syst Lab S-16440 Kista Sweden Intel Labs Neuromorph Comp Lab Santa Clara CA 95054 USA Lulea Univ Technol Dept Comp Sci Elect & Space Engn S-97187 Lulea Sweden Univ Calif Berkeley Dept Elect Engn & Comp Sci Berkeley CA 94720 USA Int Res & Training Ctr Informat Technol & Syst UA-03680 Kiev Ukraine IBM Res Zurich CH-8803 Ruschlikon Switzerland

This article reviews recent progress in the development of the computing framework vector symbolic architectures (VSA) (also known as hyperdimensional computing). This framework is well suited for implementation in stochastic, emerging hardware, and it naturally expresses the types of cognitive operations required for artificial intelligence (AI). We demonstrate in this article that the field-like algebraic structure of VSA offers simple but powerful operations on high-dimensional vectors that can support all data structures and manipulations relevant to modern computing. In addition, we illustrate the distinguishing feature of VSA, "computing in superposition," which sets it apart from conventional computing. It also opens the door to efficient solutions to the difficult combinatorial search problems inherent in AI applications. We sketch ways of demonstrating that VSA are computationally universal. We see them acting as a framework for computing with distributed representations that can play a role of an abstraction layer for emerging computing hardware. This article serves as a reference for computer architects by illustrating the philosophy behind VSA, techniques of distributed computing with them, and their relevance to emerging computing hardware, such as neuromorphic computing.

关键词： Support vector machines Philosophical considerations Neuromorphic engineering Search problems Data structures Hardware Artificial intelligence Vectors Turing completeness computing framework computing in superposition data structures distributed representations emerging hardware holographic reduced representation (HRR) hyperdimensional (HD) computing Turing completeness vector symbolic architectures (VSA)

来源：评论

学校读者我要写书评

暂无评论

Research on the computing framework in Big Data Environment 3

Research on the Computing Framework in Big Data Environment

引用

3rd International Conference on Information Science and Control Engineering (ICISCE)

作者： Liu, Yunqing Zhang, Jianhua Han, Shuqing Zhu, Mengshuai Chinese Acad Agr Sci Agr Informat Inst Beijing Peoples R China Minist Agr Agr Informat Inst CAAS Key Lab Agriinformat Serv Technol Beijing Peoples R China

ISBN: (纸本)9781509025350

computing framework is one of the key technologies in improving data analytics and processing efficiency. Since open source big data computing platform Hadoop was born ten years ago, many research achievements have been made in information acquisition, analytical processing and integrated services. Several improved frameworks were proposed against the limitations of the first generation of Map Reduce version 1 (MRv1) in scalability, reliability, efficient utilization of resource, and multiple computing model supports. This paper presents and analyzes these research results, such as batch computing framework, iterative computing framework, interactive computing framework, stream computing framework, and real-time computing framework. Undoubtedly, more targeted computing models will be generated in different application fields in the future, and these computing frameworks will play an increasingly important role in the field of big data.

关键词： c big data computing framework

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of Parallel computing framework on Embedded Multi-core Trustworthy Systems 2

Performance Analysis of Parallel Computing Framework on Embe...

引用

International Symposium on System and Software Reliability (ISSSR)

作者： Ma, Weidong Wang, Zhiying China Acad Engn Phys Comp Lab Inst Elect Engn Mianyang Peoples R China China Acad Engn Phys Syst Lab Inst Elect Engn Mianyang Peoples R China

ISBN: (纸本)9781509055630

Due to the environmental constraints, the use of high-performance embedded multi-core processor to solve the problem of scientific computing is the only way in a specified environment. This paper discusses the parallel computing framework and method in embedded multi-core system that used in safety critical system, and analyzes the performance of computation with three methods: the single thread, multi-thread and OpenMP in loongson processor, Intel Q8300 processor and I7 860 processor. The research shows that the parallel computing framework for multi-core processors can significantly improve the efficiency of program execution when the complexity of the case is not too high. The simulation test expressed that the parallel speedup is better in the loongson platform than Q8300 although the real performance of the processor is more faster than loongson.

关键词： trustworthy systems parallel computing multi-core processor computing framework

来源：评论

学校读者我要写书评

暂无评论

ENLARGE: An Efficient SNN Simulation framework on GPU Clusters

引用

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 2023年第9期34卷 2529-2540页

作者： Qu, Peng Lin, Hui Pang, Meng Liu, Xiaofei Zheng, Weimin Zhang, Youhui Tsinghua Univ Beijing Natl Res Ctr Informat Sci & Technol Dept Comp Sci & Technol Beijing 100084 Peoples R China Tsinghua Univ Dept Precis Instrument Beijing 100084 Peoples R China Yunnan Univ Sch Informat Sci & Engn Kunming 650091 Yunnan Peoples R China Zhongguancun Lab Beijing 100084 Peoples R China

Spiking Neural Networks (SNNs) are currently the most widely used computing model for neuroscience communities. There is also an increasing research interest in exploring the potential of SNN in brain-inspired computing, artificial intelligence, and other areas. As SNNs possess distinguished characteristics that originate from biological authenticity, they require dedicated simulation frameworks to achieve usability and efficiency. However, there is no widely-used, easily accessible, high performance SNN simulation framework for GPU clusters. In this paper, we propose ENLARGE, an efficient SNN simulation framework on GPU clusters. ENLARGE provides a multi-level architecture that deals with computation, communication, and synchronization hierarchically. We also propose an efficient communication method with an all-to all communication pattern. To deal with the delay of spike delivery, which is the most distinguished SNN characteristic, several delay aware optimization methods are also proposed. We further propose a multilevel workload management method. Various experiments are carried out to demonstrate the performance and scalability of the framework, as well as the effects of the optimization methods. Test results show that ENLARGE can achieve 3.17x similar to 28.12x speedup compared with the most widely used NEST simulator and 3.26 x similar to 13.57x speedup compared with the widely used NEST GPU simulator for GPU clusters.

关键词： Spiking neural network computing framework GPU cluster brain-inspired computing high performance computing

来源：评论

学校读者我要写书评

暂无评论

Application of Physics-Informed Attention-based Neural Networks for Anomaly Detection

Application of Physics-Informed Attention-based Neural Netwo...

引用

AIAA SciTech Forum

作者： Zahid, Mohammad Farooq Virginia Polytech Inst & State Univ Kevin T Crofton Dept Aerosp & Ocean Engn Blacksburg VA 24060 USA

ISBN: (数字)9781624107115

ISBN: (纸本)9781624107115

In any system, the detection of anomalous signals is of fundamental importance to ensuring reliability and efficiency. The application of novel machine learning methods allow for more prompt and informed decision making and alerts, due to the ability to model complex phenomena. While attention-based networks have enjoyed great acclaim in the language-learning space, previous work has failed to utilize advanced physics-informed attention-based networks on time-series series data with static covariates. In this work, we utilize the Temporal Fusion Transformer, which combines recurrent layers and attention layers for learning temporal relations and different time scales, with physics-informed loss functions based on the governing differential equations and static covariates for regression on the chaotic Lorenz system. We benchmark this approach on this chaotic system dataset and aim to show that this approach is viable for time-series analysis and functions well with less training than other models. We expect this approach to reduce training time for models in time-series data without any loss in performance, and thus show that production use is viable for anomaly detection in the future.

关键词： Anomaly Detection Long Short Term Memory Transformer Remaining Useful Life Catastrophic Failure Orbital Mechanics Flight Testing computing framework Natural Language Processing Quantum Mechanics

来源：评论

学校读者我要写书评

暂无评论

Novel Data-Based Model for Future Epidemiology 5

Novel Data-Based Model for Future Epidemiology

引用

5th International Conference on Artificial Intelligence, Big Data, computing and Data Communication Systems (ICABCD)

作者： Mathebula, Dephney Univ South Africa Decis Sci Pretoria South Africa

ISBN: (数字)9781665484220

ISBN: (纸本)9781665484220

Epidemiological studies aim at predicting the outbreak of diseases as epidemics and pandemics. This goal is often realized using closed form expressions that significantly utilize systems of differential equations. These systems of equations are often derived by groups of researchers working in global collaborative efforts. However, groups of researchers can experience a high workload in scenarios when there is large number of non - active participating researchers in a collaborative study. In addition, data explosion and increased availability can also undermine the efforts of hardworking researchers. This is because of the high velocity and variety associated with big data availability. Hence, an approach that helps researchers to address these challenges is required The discussion in this research proposes a suitable solution in this regard The proposed solution introduces the notion of cognitive epidemiology and epidemiological crawlers in a novel computing framework In the proposed computing framework, crawlers and existing closed form expressions used in epidemiological studies interact Prior to this interaction, epidemiological closed form expressions are formatted in a manner to enable the incorporation of intelligence capability. The research presents execution paths and discusses the incorporation alongside the integration of the proposed mechanism in a manner suitable for integration with the internet.

关键词： Artificial Intelligence Goals Epidemiology Internet Crawlers Machine learning computing framework Conventional Epidemiological Cycle COVID-19 bi-directional communications

来源：评论

学校读者我要写书评

暂无评论

High Performance Evolutionary Computation with Tensor-based Acceleration

High Performance Evolutionary Computation with Tensor-based ...

引用

Genetic and Evolutionary Computation Conference (GECCO)

作者： Klosko, Jonatan Benecki, Mateusz Wcislo, Grzegorz Dajda, Jacek Turek, Wojciech AGH Univ Sci & Technol Krakow Poland

ISBN: (纸本)9781450392372

Optimization methods, including evolutionary algorithms, need increasing efficiency in order to solve more and more complex problems in the shortest possible time. Apart from tuning the structure of those algorithms and adapting them to the given problem, the best way to speed up computations is introducing more parallelism, either at the hardware level (by using accelerators), or at the architecture level (by using multiple compute nodes). In this paper, we propose a method for expressing evolutionary computations, based on the tensor computational model. The presented approach enables cross-platform hardware acceleration on CPUs, GPUs and TPUs. To validate the new method, we contribute an open, extensible evolutionary framework, with support for distributed, accelerated execution in heterogeneous environments. Finally, we demonstrate results of the conducted tests that confirm the efficiency of the proposed approach, also in comparison to other existing frameworks.

关键词： evolutionary computing computing framework distributed computing GPGPU

来源：评论

学校读者我要写书评

暂无评论

Gene function and cell surface protein association analysis based on single-cell multiomics data

引用

COMPUTERS IN BIOLOGY AND MEDICINE 2023年第1期157卷 106733-106733页

作者： Hu, Huan Feng, Zhen Lin, Hai Cheng, Jinyan Lyu, Jie Zhang, Yaru Zhao, Junjie Xu, Fei Lin, Tao Zhao, Qi Shuai, Jianwei Xiamen Univ Dept Phys Fujian Prov Key Lab Soft Funct Mat Res Xiamen 361005 Peoples R China Xiamen Univ Natl Inst Data Sci Hlth & Med Innovat Ctr Cell Signaling Network State Key Lab Cellular Stress Biol Xiamen 361005 Peoples R China Univ Chinese Acad Sci Wenzhou Inst Wenzhou 325001 Peoples R China Univ Chinese Acad Sci Wenzhou Key Lab Biophys Wenzhou 325001 Peoples R China Wenzhou Med Univ Affiliated Hosp 1 Wenzhou 325000 Peoples R China Wenzhou Med Univ Sch Biomed Engn Inst Biomed Big Data Sch Ophthalmol & Optometry Wenzhou 325027 Peoples R China Wenzhou Med Univ Eye Hosp Sch Biomed Engn Wenzhou 325027 Peoples R China Guangzhou Univ Cyberspace Inst Adv Technol Guangzhou 510000 Peoples R China Zhejiang Lab Regenerat Med Vis & Brain Hlth Oujiang Lab Wenzhou 325001 Peoples R China Univ Sci & Technol Liaoning Sch Comp Sci & Software Engn Anshan 114051 Peoples R China Xiamen Univ Dept Phys Xiamen 361005 Peoples R China Xiamen Univ Fujian Prov Key Lab Soft Funct Mat Res Xiamen 361005 Peoples R China

Single-cell transcriptomics provides researchers with a powerful tool to resolve the transcriptome heterogeneity of individual cells. However, this method falls short in revealing cellular heterogeneity at the protein level. Previous single-cell multiomics studies have focused on data integration rather than exploiting the full potential of multiomics data. Here we introduce a new analysis framework, gene function and protein association (GFPA), that mines reliable associations between gene function and cell surface protein from single-cell multimodal data. Applying GFPA to human peripheral blood mononuclear cells (PBMCs), we observe an association of epithelial mesenchymal transition (EMT) with the CD99 protein in CD4 T cells, which is consistent with previous findings. Our results show that GFPA is reliable across multiple cell subtypes and PBMC samples. The GFPA python packages and detailed tutorials are freely available at https://***/studentiz/GFPA.

关键词： Single-cell Multiomics Cell surface protein Association analysis computing framework

来源：评论

学校读者我要写书评

暂无评论

Data-Driven Modeling of Dynamic Systems Based on Online Learning

Data-Driven Modeling of Dynamic Systems Based on Online Lear...

引用

AIAA Propulsion and Energy Forum / IEEE Electric Aircraft Technologies Symposium (EATS)

作者： Jiang, Zhenhua Beigh, Kelly Univ Dayton Res Inst Dayton OH 45469 USA

ISBN: (数字)9781624106118

ISBN: (纸本)9781624106118

This paper presents modeling and prediction of the unknown dynamics from a variety of real-time operation data in complex physical processes or systems based on machine learning. More generally, this is a data-driven modeling approach based on physical knowledge at time of design and machine learning at runtime. The learning process may apply to the entire or partial system dynamics, while portions of the physical system dynamics may be known or given. The learned models together with the known physics-based equations can be widely used in control system design, for instance, as one component in a model predictive control (MPC) framework for integrated dynamic systems. Long short-term memory (LSTM) neural networks are chosen here, because they have additional stored states resulting from the past output, and the state storage can be internally controlled subject to the network status itself. Such controlled states can be regarded as gated states or memory blocks in a neural network, and they serve as key components of the LSTM neural network in controlling the information flow throughout the network. This paper provides technical details about the concept of a LSTM neural network that can be used to model and predict energy production in renewable sources, battery state of charge, engine fuel use (or energy efficiency), or energy demand of the payload in aerospace or energy system applications. The most compute-intensive part of the learning algorithm is solving a nonlinear optimization problem to minimize the total error between the predicted and actual outputs. After updating the learned model and parameters, the LSTM model implementation may perform certain computational steps in parallel, so as to accelerate the computing. The model learning and prediction processes can be implemented on a field programmable gate array (FPGA - a reconfigurable parallel computing device), for instance, on NI's CompactRIO modules or PXI FPGA modules. As an example, an application

关键词： Feedforward Neural Network Field Programmable Gate Array Machine Learning Energy Production Model Predictive Control computing framework Control System Design Energy Efficiency Parallel computing Energy Systems

来源：评论

学校读者我要写书评

暂无评论

Multi-Fidelity, Gradient-enhanced, and Locally Optimized Sparse Polynomial Chaos and Kriging Surrogate Models Applied to Benchmark Problems

Multi-Fidelity, Gradient-enhanced, and Locally Optimized Spa...

引用

AIAA Science and Technology Forum and Exposition (AIAA SciTech)

作者： Rumpfkeil, Markus P. Beran, Philip S. Univ Dayton Dept Mech & Aerosp Engn Dayton OH 45469 USA US Air Force Res Lab Wright Patterson AFB OH 45433 USA

ISBN: (数字)9781624105951

ISBN: (纸本)9781624105951

In this paper, multi-fidelity, gradient-enhanced, and locally optimized sparse polynomial chaos expansion (PCE) and kriging surrogate models are constructed in lieu of solely using computationally expensive high-fidelity engineering analyses. Once an accurate surrogate model is built, it can be used for evaluating a large number of designs for design space exploration or for uncertainty quantification. To demonstrate that accurate multi-fidelity, gradient-enhanced, and locally optimized surrogate models can be obtained at lower computational cost than basic high-fidelity ones, a number of benchmark problems are employed. These include polynomial and non-polynomial analytical functions of arbitrary dimension, a heterogeneous non-polynomial analytical function in one, two, and three dimensions as well as a coupled spring-mass-system leading to a linear system of ODEs of arbitrary dimension.

关键词： Surrogate Model Space Exploration Probability Distribution computing framework Air Force Research Laboratory Singular Value Decomposition Gaussian Mixture Models Mass Spring System Stiffness Matrices Maximum Likelihood Estimation

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：