495 Hits in 2.5 sec

A Survey of Machine Learning-Based System Performance Optimization Techniques

Hyejeong Choi, Sejin Park
2021 Applied Sciences  
This survey provides a detailed design and summarizes model, input, output, and prediction method of each approach.  ...  This paper covers various system performance areas from the data structure to essential system components of a computer system such as index structure, branch predictor, sort, and cache management.  ...  Data Availability Statement: Not applicable. Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/app11073235 fatcat:fjjguf3x6fdjtm543iejmr255i

Learning Memory Access Patterns [article]

Milad Hashemi, Kevin Swersky, Jamie A. Smith, Grant Ayers, Heiner Litz, Jichuan Chang, Christos Kozyrakis, Parthasarathy Ranganathan
2018 arXiv   pre-print
We relate contemporary prefetching strategies to n-gram models in natural language processing, and show how recurrent neural networks can serve as a drop-in replacement.  ...  This work represents the first step towards practical neural-network based prefetching, and opens a wide range of exciting directions for machine learning in computer architecture research.  ...  Our clustering approach is also reminiscent of the hierarchical approach that they deploy. Importantly, they find that neural network models are faster to query than conventional data structures.  ... 
arXiv:1803.02329v1 fatcat:35topihcwvgdtc2udajbkptbmi

Neural Network Based Prefetching Control Mechanism

2019 International Journal of Engineering and Advanced Technology  
This paper is an attempt to dynamically monitor the network bandwidth for which a neural network-based model has been worked upon.  ...  Therefore, there is critical need of a mechanism that could analyze the network bandwidth of the system before prefetching is done.  ...  In the next subsection, we present neural network basics and neural network based proposed model. A. Neural Network A Neural Network is a collection of artificial neurons.  ... 
doi:10.35940/ijeat.b2621.129219 fatcat:6prf5hmws5eztoy72x75n6jvwm

Prefetching of VoD Programs Based On ART1 Requesting Clustering [article]

P. Jayarekha, T. R. GopalaKrishnan Nair
2009 arXiv   pre-print
The simulation results of our proposed clustering and prefetching algorithm, shows enormous increase in the performance of streaming server.  ...  In this paper, we propose a novel approach to group users according to the VoD user request pattern. We cluster the user requests based on ART1 neural network algorithm.  ...  A target is set as the request arrival pattern, to achieve the results comparative with that of target the videos are prefetched using a ART1 model for which a set of videos to be prefetched are grouped  ... 
arXiv:0910.1468v1 fatcat:mrwqe6z24bga7je2usxwvzthdi

Adaptive neural network clustering of Web users

S.K. Rangarajan, V.V. Phoha, K.S. Balagani, R.R. Selmic, S.S. Iyengar
2004 Computer  
A prefetching application of this clustering technique showed prediction accuracy as high as 97.78 percent.  ...  A neural network based on adaptive resonance theory dynamically groups users based on their Web access patterns.  ...  Hierarchical clustering is a statistical method for finding clusters of identical data points.  ... 
doi:10.1109/mc.2004.1297299 fatcat:mt4iivfj7fdhxenokh2utbn4zu

Functionality-Based Processing-In-Memory Accelerator for Deep Convolutional Neural Networks

Min-Jae Kim, Jeong-Geun Kim, Su-Kyung Yoon, Shin-Dug Kim
2021 IEEE Access  
In particular, deep convolutional neural networks (DCNNs) processing that consists of several functionalities could be highly optimized if PIM cores can extend the processing capability and data accessibility  ...  Second, an efficient replacement method complements the shared cache to optimize the data miss rate of DCNN processing.  ...  Additionally, we found that our shared cache model (column 3) showed a slightly lower bandwidth with an average of 83.5 GB/sec, than our dual prefetchers model (column 2), which has an average of 84.8  ... 
doi:10.1109/access.2021.3122818 fatcat:eer2kddywje7fjugcud5uokulu

Learning Execution through Neural Code Fusion [article]

Zhan Shi, Kevin Swersky, Daniel Tarlow, Parthasarathy Ranganathan, Milad Hashemi
2020 arXiv   pre-print
data structures into a simpler, more uniform format.  ...  While there is a growing body of work on using Graph Neural Networks (GNNs) to learn representations of source code, these representations do not understand how code dynamically executes.  ...  To our knowledge, NCF is the first instance of a single model that can learn simultaneously on dynamic control-flow and data-flow tasks, setting the stage for teaching neural network models to better understand  ... 
arXiv:1906.07181v2 fatcat:czxoxlbifbgn3kvssezsyzjclu

A Generic Architecture for Hybrid Intelligent Systems [chapter]

Hans-Arno Jacobsen
2001 Studies in Fuzziness and Soft Computing  
Two original instantiations of this framework are presented and discussed. Their performance is evaluated for prefetching of bulk data over wireless media.  ...  We think it is crucial for the design of intelligent systems to focus on the integration and interaction of di erent learning techniques in one model rather then merging them to create ever new techniques  ...  Acknowledgements The author is very grateful to the following people for assisting in one way or another in the preparation of this paper: Lot Zadeh, Shun'ichi Tano, and Joachim Weisbrod.  ... 
doi:10.1007/978-3-7908-1837-6_7 fatcat:ugrut7lv7bgq3nmpgyaxacba74

ARCHIE: Data Analysis Acceleration with Array Caching in Hierarchical Storage

Bin Dong, Teng Wang, Houjun Tang, Quincey Koziol, Kesheng Wu, Suren Byna
2018 2018 IEEE International Conference on Big Data (Big Data)  
In this paper, we introduce a new array caching in hierarchical storage (ARCHIE) to accelerate array data analyses in a seamless fashion.  ...  ARCHIE evaluates array access patterns and prefetches data with array semantics between storage layers.  ...  With the goal of providing a transparent and efficient data prefetching solution using array semantics in a hierarchical storage subsystem, we propose a Array Caching in HIErarchical storage system (ARCHIE  ... 
doi:10.1109/bigdata.2018.8622616 dblp:conf/bigdataconf/DongWTKWB18 fatcat:tqb65jrrhnenjpe7ahc2omgq74

Local Citation Recommendation with Hierarchical-Attention Text Encoder and SciBERT-based Reranking [article]

Nianlong Gu, Yingqiang Gao, Richard H.R. Hahnloser
2022 arXiv   pre-print
When coupled with a SciBERT reranker fine-tuned on local citation recommendation tasks, our hierarchical Attention encoder (HAtten) achieves high prefetch recall for a given number of candidates to be  ...  To balance the tradeoff between speed and accuracy of citation recommendation in the context of a large-scale paper database, a viable approach is to first prefetch a limited number of relevant documents  ...  in the dataset, which eliminates a large bulk of data.  ... 
arXiv:2112.01206v3 fatcat:4vsjgxiqfbgadcmarn7dzwvsea

Enhancement of Performance of Proxy Server by Reducing Web Traffic using Web Usage Mining

2015 International Journal of Science and Research (IJSR)  
., downloading) a Web request before the user actually makes it. By doing so, the waiting time perceived by the user can be reduced, which is the main goal of the Web prefetching techniques.  ...  The study of the state of the art about Web prefetching showed the heterogeneity that exists in its performance evaluation.  ...  A scheme for fast allocation of web pages using data mining techniques and competitive neural network is being discussed in.  ... 
doi:10.21275/v5i5.nov163345 fatcat:nulc5xj43jecxfhn3qug2ab7ua

A Novel Memory-Scheduling Strategy for Large Convolutional Neural Network on Memory-Limited Devices

Shijie Li, Xiaolong Shen, Yong Dou, Shice Ni, Jinwei Xu, Ke Yang, Qiang Wang, Xin Niu
2019 Computational Intelligence and Neuroscience  
However, traditional deep learning methods such as convolutional neural network (CNN) and its variants consume a lot of memory resources.  ...  neural networks).  ...  data of the current layer from the network model.  ... 
doi:10.1155/2019/4328653 pmid:31182958 pmcid:PMC6512078 fatcat:tjihs56jujez5ntmzoaxw62cwi

A Survey of Machine Learning Applied to Computer Architecture Design [article]

Drew D. Penney, Lizhong Chen
2019 arXiv   pre-print
Notably, machine learning based strategies often surpass prior state-of-the-art analytical, heuristic, and human-expert approaches.  ...  Taken together, these strategies and techniques present a promising future for increasingly automated architectural design.  ...  Zeng and Guo [26] proposed a long short-term memory (LSTM) model (a recurrent neural network variant) for prefetching based on local history and offset-delta tables.  ... 
arXiv:1909.12373v1 fatcat:o4nscgkjfbes7kqwmtjvvgl3oa

Distributed Hierarchical GPU Parameter Server for Massive Scale Deep Learning Ads Systems [article]

Weijie Zhao, Deping Xie, Ronglai Jia, Yulei Qian, Ruiquan Ding, Mingming Sun, Ping Li
2020 arXiv   pre-print
All the neural network training computations are contained in GPUs. Extensive experiments on real-world data confirm the effectiveness and the scalability of the proposed system.  ...  A 4-node hierarchical GPU parameter server can train a model more than 2X faster than a 150-node in-memory distributed parameter server in an MPI cluster.  ...  We build a 4-stage pipeline to hide the latency of those tasks by maintaining a prefetch queue for each stage.  ... 
arXiv:2003.05622v1 fatcat:kfl2uv7oarfsfa7zpkgps76h6e

Web page access prediction using hierarchical clustering based on modified levenshtein distance and higher order Markov model

B T Harish Kumar, L Vibha, K R Venugopal
2016 2016 IEEE Region 10 Symposium (TENSYMP)  
Predictions need to keep track of history data to analyze the usage behavior of the users. Web Usage behavior of a user can be analyzed using the web log file of a specific website.  ...  The proposed research work uses hierarchical clustering technique with modified Levenshtein distance, Page Rank using access time length, frequency and higher order Markov model for prediction.  ...  Hierarchical Clustering Hierarchical Clustering is a method in data mining which allows clustering analysis.  ... 
doi:10.1109/tenconspring.2016.7519368 fatcat:ickly6g6drez3ettu4fr4d2nwu
« Previous Showing results 1 — 15 out of 495 results