检索结果-内蒙古大学图书馆

Exploring Auditory-Inspired Acoustic Features for Room Acoustic Parameter Estimation From Monaural Speech

ieee-acm TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING 2018年第10期26卷 1809-1820页

作者： Xiong, Feifei Goetze, Stefan Kollmeier, Birger Meyer, Bernd T. Carl von Ossietzky Univ Oldenburg Med Phys Dept D-26129 Oldenburg Germany Carl von Ossietzky Univ Oldenburg Cluster Excellence Hearing4all D-26129 Oldenburg Germany Fraunhofer Inst Digital Media Technol IDMT D-26129 Oldenburg Germany Cluster Excellence Hearing4all D-26129 Oldenburg Germany

Room acoustic parameters that characterize acoustic environments can help to improve signal enhancement algorithms such as for dereverberation, or automatic speech recognition by adapting models to the current parameter set. The reverberation time (RT) and the early-to-late reverberation ratio (ELR) are two key parameters. In this paper, we propose a blind ROom Parameter Estimator (ROPE) based on an artificial neural network that learns the mapping to discrete ranges of the RT and the ELR from single-microphone speech signals. Auditory-inspired acoustic features are used as neural network input, which are generated by a temporal modulation filter bank applied to the speech time-frequency representation. ROPE performance is analyzed in various reverberant environments in both clean and noisy conditions for both fullband and subband RT and ELR estimations. The importance of specific temporal modulation frequencies is analyzed by evaluating the contribution of individual filters to the ROPE performance. Experimental results show that ROPE is robust against different variations caused by room impulse responses (measured versus simulated), mismatched noise levels, and speech variability reflected through different corpora. Compared to state-of-the-art algorithms that were tested in the acoustic characterisation of environments (ACE) challenge, the ROPE model is the only one that is among the best for all individual tasks (RT and ELR estimation from fullband and subband signals). Improved fullband estimations are even obtained by ROPE when integrating speech-related frequency subbands. Furthermore, the model requires the least computational resources with a real time factor that is at least two times faster than competing algorithms. Results are achieved with an average observation window of 3 s, which is important for real-time applications.

关键词： Reverberation time early-to-late reverberation ratio blind estimation auditory-inspired acoustic features machine learning

来源：评论

学校读者我要写书评

暂无评论

All Eyes on You: Distributed Multi-Dimensional IoT Microservice Anomaly Detection 14

All Eyes on You: Distributed Multi-Dimensional IoT Microserv...

引用

14th International Conference on Network and Service Management (CNSM)

作者： Pahl, Marc-Oliver Aubet, Francois-Xavier Tech Univ Munich Munich Germany

ISBN: (纸本)9783903176140

The Internet of Things (IoT) is a Distributed System of cooperating Microservices (mu Ss). IoT services manage devices that monitor and control their environments. The interaction of the IoT with the physical environment creates strong security, privacy, and safety implications. It makes providing adequate security for IoT mu Ss essential. However, the complexity of IoT services makes detecting anomalous behavior difficult. We present a machine-learning based approach for modeling IoT service behavior by only observing inter-service communication. Our algorithm continuously learns S models on distributed IoT nodes within an IoT site. Combining the learned models within and in-between IoT sites converges our mu S models within short time. Sharing the resulting stable models among compute nodes enables good anomaly detection. As one application, firewalling IoT mu Ss becomes possible. Combining our autonomous mu S modeling with firewalling enables retrofitting security to existing IoT installations. We enable retrofitting access control to existing non-secure IoT installations. Our proposed approach is resource efficient, matching the requirements of the IoT. To evaluate the quality of our proposed algorithm, we show its behavior for a set of common IoT attacks. We evaluate how domain knowledge enables us to decorrelate events on a node, and how adding context features improves the detection rate.

关键词： IoT security machine learning modeling anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Data Placement Optimization in GPU Memory Hierarchy using Predictive Modeling

Data Placement Optimization in GPU Memory Hierarchy using Pr...

引用

Workshop on Memory Centric High Performance Computing (MChpc)

作者： Stoltzfus, Larisa Emani, Murali Lin, Pei-Hung Liao, Chunhua Univ Edinburgh Edinburgh Midlothian Scotland Lawrence Livermore Natl Lab Livermore CA 94550 USA

ISBN: (纸本)9781450361132

Modern supercomputers often use Graphic Processing Units (or GPUs) to meet the ever-growing demands for high performance computing. GPUs typically have a complex memory architecture with various types of memories and caches, such as global memory, shared memory, constant memory, and texture memory. The placement of data on these memories has a tremendous impact on the performance of the hpc applications and identifying the optimal placement location is non-trivial. In this paper, we propose a machine learning-based approach to build a classifier to determine the best class of GPU memory that will minimize GPU kernel execution time. This approach utilizes a set of performance counters obtained from profiling runs along with hardware features to generate the trained model. We evaluate our approach on several generations of NVIDIA GPUs, including Kepler, Maxwell, Pascal, and Volta on a set of benchmarks. The results show that the trained model achieves prediction accuracy over 90% and given a global version, the classifier can accurately determine which data placement variant would yield the best performance.

关键词： GPU Data placement Memory machine learning

来源：评论

学校读者我要写书评

暂无评论

Prediction and Modeling for No-Reference Video Quality Assessment based on machine learning 14

Prediction and Modeling for No-Reference Video Quality Asses...

引用

14th International Conference on Signal Image Technology & Internet Based Systems (SITIS)

作者： Pedro Lopez, Juan Martin, David Jimenez, David Manuel Menendez, Jose Univ Politecn Madrid Signals Syst & Radiocommun Dept Madrid Spain

ISBN: (纸本)9781538693858

The increase in popularity of video streaming, the improvement of bandwidth corresponding to 5G networks and the transmission of higher amounts of data derived from advanced video formats such as Ultra High Definition (UHD) with 4K and 8K resolutions make the user demand a high perceptual quality of the contents consumed. For that reason, it is necessary to create advanced models for predicting video quality based on the video features and the encoding settings in environments where the reference is not present. The use of machine learning (ML) techniques for data analysis based on patterns extracted on the features of audio-visual content, improves the generation of prediction models for accurately predicting quality. This paper presents a novel model for assessing video quality based on the analysis of encoding video settings of the transmitted contents and the image intrinsic characteristics for objectively estimating the Mean Opinion Score (MOS) in correlation with the subjective results. The use of data mining techniques for combining a collection of parameters associated to the video transmitted improves the performance of traditional quality evaluation, as demonstrated with the database analyzed for this purpose.

关键词： Video quality assessment machine-learning MOS estimation prediction models video features multimedia encoding techniques

来源：评论

学校读者我要写书评

暂无评论

Across the Stack Opportunities for Deep learning Acceleration 18

Across the Stack Opportunities for Deep Learning Acceleratio...

引用

23rd ieee/acm International Symposium on Low Power Electronics and Design (ISLPED)

作者： Srinivasan, Vijayalakshmi Fleischer, Bruce Shukla, Sunil Ziegler, Matthew Silberman, Joel Oh, Jinwook Choi, Jungwook Mueller, Silvia Agrawal, Ankur Babinsky, Tina Cao, Nianzheng Chen, Chia-Yu Chuang, Pierce Fox, Thomas Gristede, George Guillorn, Michael Haynie, Howard Klaiber, Michael Lee, Dongsoo Lo, Shih-Hsien Maier, Gary Scheuermann, Michael Venkataramani, Swagath Vezyrtzis, Christos Wang, Naigang Yee, Fanchieh Zhou, Ching Lu, Pong-Fei Curran, Brian Chang, Leland Gopalakrishnan, Kailash IBM TJ Watson Res Ctr Yorktown Hts NY 10598 USA IBM Syst Grp Poughkeepsie NY USA

ISBN: (纸本)9781450357043

The combination of growth in compute capabilities and availability of large datasets has led to a re-birth of deep learning. Deep Neural Networks (DNNs) have become state-of-the-art in a variety of machine learning tasks spanning domains across vision, speech, and machine translation. Deep learning (DL) achieves high accuracy in these tasks at the expense of 100s of ExaOps of computation; posing significant challenges to efficient large-scale deployment in both resource-constrained environments and data *** of the key enablers to improve operational efficiency of DNNs is the observation that when extracting deep insight from vast quantities of structured and unstructured data the exactness imposed by traditional computing is not required. Relaxing the "exactness" constraint enables exploiting opportunities for approximate computing across all layers of the system *** this talk we present a multi-TOPS AI core [3] for acceleration of deep learning training and inference in systems from edge devices to data centers. We demonstrate that to derive high sustained utilization and energy efficiency from the AI core requires ground-up re-thinking to exploit approximate computing across the stack including algorithms, architecture, programmability, and *** accuracy is the fundamental measure of deep learning quality. The compute engine precision in our AI core is carefully calibrated to realize significant reduction in area and power while not compromising numerical accuracy. Our research at the DL algorithms/applications-level [2] shows that it is possible to carefully tune the precision of both weights and activations to as low as 2-bits for inference and was used to guide the choices of compute precision supported in the architecture and hardware for both training and inference. Similarly, distributed DL training's scalability is impacted by the communication overhead to exchange gradients and weights after each mini-batch. Our research on gradient

关键词： Deep learning Accelerators DNN Optimizations DL Training DL inference

来源：评论

学校读者我要写书评

暂无评论

NUMA-Caffe: NUMA-Aware Deep learning Neural Networks

引用

acm TRANSACTIONS ON ARCHITECTURE AND CODE OPTIMIZATION 2018年第2期15卷 24-24页

作者： Roy, Probir Song, Shuaiwen Leon Krishnamoorthy, Sriram Vishnu, Abhinav Sengupta, Dipanjan Liu, Xu Coll William & Mary Dept Comp Sci McGlothlin St Hall 117 Williamsburg VA 23185 USA Pacific Northwest Natl Lab High Performance Comp Grp Richland WA 99352 USA Coll William & Mary Williamsburg VA 23187 USA Intel Labs Parallel Comp Lab 2200 Mission Coll Blvd Santa Clara CA 95054 USA

Convolution Neural Networks (CNNs), a special subcategory of Deep learning Neural Networks (DNNs), have become increasingly popular in industry and academia for their powerful capability in pattern classification, image processing, and speech recognition. Recently, they have been widely adopted in High Performance Computing (hpc) environments for solving complex problems related to modeling, runtime prediction, and big data analysis. Current state-of-the-art designs for DNNs on modern multi- and many-core CPU architectures, such as variants of Caffe, have reported promising performance in speedup and scalability, comparable with the GPU implementations. However, modern CPU architectures employ Non-Uniform Memory Access (NUMA) technique to integrate multiple sockets, which incurs unique challenges for designing highly efficient CNN frameworks. Without a careful design, DNN frameworks can easily suffer from long memory latency due to a large number of memory accesses to remote NUMA domains, resulting in poor scalability. To address this challenge, we propose NUMA-aware multi-solver-based CNN design, named NUMA-Caffe, for accelerating deep learning neural networks on multi- and many-core CPU architectures. NUMA-Caffe is independent of DNN topology, does not impact network convergence rates, and provides superior scalability to the existing Caffe variants. Through a thorough empirical study on four contemporary NUMA-based multi- and many-core architectures, our experimental results demonstrate that NUMA-Caffe significantly outperforms the state-of-the-art Caffe designs in terms of both throughput and scalability.

关键词： Deep learning neural network NUMA stochastic gradient descent

来源：评论

学校读者我要写书评

暂无评论

Evolving deep networks using hpc

Evolving deep networks using HPC

引用

2017 machine learning in hpc environments, mlhpc 2017

作者： Young, Steven R. Rose, Derek C. Johnston, Travis Heller, William T. Karnowski, Thomas P. Potok, Thomas E. Patton, Robert M. Perdue, Gabriel Miller, Jonathan Oak Ridge National Laboratory Oak Ridge United States Fermi National Accelerator Laboratory BataviaIL United States Universidad Técnica Federico Santa María Avenida España Valparaíso Chile

ISBN: (纸本)9781450351379

While a large number of deep learning networks have been studied and published that produce outstanding results on natural image datasets, these datasets only make up a fraction of those to which deep learning can be applied. These datasets include text data, audio data, and arrays of sensors that have very different characteristics than natural images. As these "best" networks for natural images have been largely discovered through experimentation and cannot be proven optimal on some theoretical basis, there is no reason to believe that they are the optimal network for these drastically different datasets. Hyperparameter search is thus often a very important process when applying deep learning to a new problem. In this work we present an evolutionary approach to searching the possible space of network hyperparameters and construction that can scale to 18, 000 nodes. This approach is applied to datasets of varying types and characteristics where we demonstrate the ability to rapidly find best hyperparameters in order to enable practitioners to quickly iterate between idea and result. © 2017 Copyright held by the owner/author(s).

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Proceedings of mlhpc 2015: machine learning in High-Performance Computing environments - Held in conjunction with SC 2015: The International Conference for High Performance Computing, Networking, Storage and Analysis

Proceedings of MLHPC 2015: Machine Learning in High-Performa...

引用

Workshop on machine learning in High-Performance Computing environments, mlhpc 2015

ISBN: (纸本)9781450340069

The proceedings contain 5 papers. The topics discussed include: asynchronous parallel stochastic gradient descent - a numeric core for scalable distributed machine learning algorithms;HPDBSCAN - highly parallel DBSCAN;dynamic adaptive neural network arrays: a neuromorphic architecture;optimizing deep learning hyper-parameters through an evolutionary algorithm;and LBANN: Livermore big artificial neural network hpc toolkit.

关键词：

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：