Yuta Inouchi, Hayato Yamaki, Shinobu Miwa, Tomoaki Tsumura
{"title":"Functionally-Predefined Kernel: a Way to Reduce CNN Computation","authors":"Yuta Inouchi, Hayato Yamaki, Shinobu Miwa, Tomoaki Tsumura","doi":"10.1109/PACRIM47961.2019.8985122","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985122","url":null,"abstract":"Convolutional Neural Networks (CNNs) have achieved high classification accuracy in image recognition, and now, they are widely used for numerous applications. For higher accuracy or more advanced applications, CNNs need to consume tremendous computational resources and time. Hence, many studies for reducing the computational cost of CNNs are actively being conducted. However, many previous methods for reducing the computational cost lead to a non-negligible loss in output accuracy. Therefore, it is still a challenge to reduce the computational cost of CNNs with keeping the output accuracy high. In this paper, we propose a novel concept \"Functionally-Predefined Kernel\" to reduce the computational cost for CNN training and discuss the potential of computation reuse to reduce the computational cost for CNN inference. Our experimental results show that the number of parameters to be trained can be significantly reduced by utilizing Functionally-Predefined Kernels without accuracy loss. In addition, we revealed that CNN’s inference process includes many convolution operations with the same inputs and computation reuse, therefore, has high affinity to CNN computation.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124426472","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Programmability and Performance of New Global-View Programming API for Multi-Node and Multi-Core Processing","authors":"Yugo Sakaguchi, H. Midorikawa","doi":"10.1109/PACRIM47961.2019.8985126","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985126","url":null,"abstract":"Various partitioned global address space (PGAS) languages capable of providing global-view programming environments on multi-node computer systems have been proposed to improve programming productivity in high-performance computing. However, several PGAS languages often require a detailed description of the remote data access, similar to descriptions used in message passing interface one-sided communications. Some PGAS languages have limitations pertaining to remote data access and recommend their local-view programming models, rather than the global-view ones, due to performance-related reasons. In this study, we propose SMint, which is an application programming interface that provides a global-view programming model with a software distributed shared memory mSMS as the runtime. Using stencil computation as a typical processing method, the performance and programmability of SMint have been compared with those of XcalableMP and Unified Parallel C, which are well-known examples of PGAS languages based on the C language. It was found that SMint achieved the best performance under the ideal global-view programming model.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"159 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114433832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Mobile User-Activity Prediction Utilizing LSTM Recurrent Neural Network","authors":"R. Sharifi, Mahdiyar Molahasani, V. Vakili","doi":"10.1109/PACRIM47961.2019.8985068","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985068","url":null,"abstract":"The demand for mobile services is increasing exponentially and mobile telecommunication operators are facing new challenges such as bandwidth choking and resource allocation issues consequently. One of the most promising methods for overcoming these challenges is using Big Data Analytics on Call Detail Record (CDR). In this paper, a Long-Short Term Memory neural network is utilized for user activity prediction for the first time. The real CDR data is preprocessed and used for training the neural network. The accuracy of this prediction is evaluated in normal values and anomalies. The proposed system can predict user activity accurately using a limited amount of training data and its performance in anomalous behaviors is promising. The networks ability to be generalized is also evaluated with the cross-dataset test. The neural network is trained with one cells data and predict another cells activity properly. The proposed system is a step forward toward designing a practical and efficient CDR anomaly prediction system.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":" 61","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113948627","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Hybrid Method for Spectral Translation Equivalent Boolean Functions","authors":"Mathias Soeken, Eleonora Testa, D. M. Miller","doi":"10.1109/PACRIM47961.2019.8985048","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985048","url":null,"abstract":"The equivalence of Boolean functions with respect to five invariance (aka translation) operations has been well considered with respect to the Rademacher-Walsh spectral domain. In this paper, we introduce a hybrid approach that uses both the Reed-Muller and the Rademacher-Walsh spectra. A novel hybrid algorithm that maps a Boolean function to a representative function for the equivalence class containing the original function is presented. The algorithm can be used to determine a sequence of translations that maps one function to an equivalent function. We present experimental results that show the hybrid algorithm can determine the equivalence classes for 5 variables much more efficiently than before. We also show that for 6 variables where there are 150,357 equivalence classes, 8 are very difficult, a further 58 are difficult and the remainder are straightforward in terms of the CPU time required by the hybrid algorithm.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114861516","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Phoneme Sequence Extraction of Learner’s Pronunciation Errors for Foreign Language Learning","authors":"Kohei Kamimura, Kosuke Takano","doi":"10.1109/PACRIM47961.2019.8985130","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985130","url":null,"abstract":"This paper presents a method for extracting phoneme sequences in foreign language pronunciation learning that a learner is likely to mispronounce. Some conventional methods show a learner’s phoneme errors using speech recognition technology. Those systems simply show only one phoneme error that the learner makes; they show no analytical results of phoneme errors that arise from previous or later phoneme sequences. A feature of the proposed method is the ability to extract a phoneme that a learner is likely to mispronounce as an n-gram sequence rather than a single phoneme. We confirm the feasibility of our method with several experiments using a prototype of our learning system.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127769650","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Timing Synchronization for Upstream Cable OFDMA Signals","authors":"Yayi Xiao, B. Berscheid, Ha H. Nguyen, J. Salt","doi":"10.1109/PACRIM47961.2019.8985094","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985094","url":null,"abstract":"The most recent broadband data-over-cable standard, known as DOCSIS 3.1, introduced orthogonal frequency-division multiplexing (OFDM) as a major upgrade to improve transmission efficiency. Because OFDM signals are sensitive to synchronization errors, accurate timing recovery is an integral part of a DOCSIS 3.1 system. Many conventional OFDM/OFDMA timing synchronization methods cannot be directly applied, as they require the use of specific training signals which are incompatible with DOCSIS 3.1. In this paper, the performance of a modified Schmidl and Cox (S&C) timing algorithm is first investigated for DOCSIS 3.1 systems and shown to be unsatisfactory under severe channel conditions. Two novel timing algorithms are proposed which take advantage of the mirrored symmetry of the training signal defined in DOCSIS 3.1. It is demonstrated that both algorithms provide accurate timing estimates even under severe channel conditions and heavy traffic flow. The second algorithm is introduced to eliminate the use of multipliers in the first algorithm, hence reducing the hardware complexity. Data skipping and data truncation are also presented as methods of further reducing the hardware complexity and the corresponding cost-performance tradeoff is investigated.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128114589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Anomaly Detection by Monitoring Unintended DNS Traffic on Wireless Network","authors":"Yong Jin, M. Tomoishi, N. Yamai","doi":"10.1109/PACRIM47961.2019.8985052","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985052","url":null,"abstract":"Cybersecurity threats from malware attacks has become one of the most serious issues in the Internet nowadays. Most types of malware, after intruding an individual computer, attempt to connect the corresponding Command and Control (C&C) servers using IP addresses or Fully Qualified Domain Name (FQDN) and receive further instructions (e.g. attacking target IP addresses and FQDNs) from them in order to conduct subsequent cyber attacks. In recent years, it has been clarified that DNS traffic has been used for communication between the malware infection computers and the C&C servers. In this research, we focus on these peculiarities and propose a method for detecting malware infected computers by monitoring unintended DNS traffic on wireless networks by collaboration with DHCP (Dynamic Host Configuration Protocol) server. By deploying the proposed system on campus wireless networks, the computers within DHCP configured environment can be detected when they are infected by some types of malware and it attempts to communicate with the corresponding C&C servers using DNS (Domain Name System) protocol. In this paper, we describe the detailed design of the proposed method and the future work includes prototype implementation as well as evaluations.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125786199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Receiver Design and SER Analysis of Massive MIMO Uplink With Mixed-Resolution ADCs","authors":"Ke Ma, Y. Jing","doi":"10.1109/PACRIM47961.2019.8985121","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985121","url":null,"abstract":"This paper is on the uplink communications of single-cell multi-user massive multi-input-multi-output (MIMO) systems with a mixture of 1-bit analog-to-digital converters (ADCs) and full-resolution ADCs. A new low-complexity linear combining scheme is proposed for the receiver at the base station (BS) to take into account the difference of the two ADC types. A symbol-wise decoding is derived to separately decode the information of each user and an analytical approximation is obtained for the symbol error rate (SER) of the proposed combining and decoding. Simulation results show that the proposed scheme achieves significant performance improvement compared to the traditional maximum-ratio-combining (MRC) receiver.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127910605","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"PACRIM 2019 Copyright Page","authors":"","doi":"10.1109/pacrim47961.2019.8985027","DOIUrl":"https://doi.org/10.1109/pacrim47961.2019.8985027","url":null,"abstract":"","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132381332","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Design of Optimal Finite Alphabet NOMA Scheme for Uplink Noncoherent Massive MIMO Channels","authors":"Yang Yu, Peiyao Chen, Jian-Kang Zhang","doi":"10.1109/PACRIM47961.2019.8985128","DOIUrl":"https://doi.org/10.1109/PACRIM47961.2019.8985128","url":null,"abstract":"The main focus of this paper is on an uplink communication system with multiple single-antenna users utilizing the same frequency band and simultaneously transmitting their finite alphabet signals to one base station having a large number of antennas. It is assumed that small scale channel fading is Rayleigh fading and varies in every one time slot, and that the large scale fading coefficients are fixed and known at both the transmitter and the receiver. For such uplink non-orthogonal multiple access massive MIMO system, we aim at establishing a new criterion using a noncoherent maximum likelihood detector for collaboratively designing all such user constellations that once each received sum signal point has been recovered, all the user signal points can be uniquely determined, i.e., all the user constellations form a so called additively uniquely decomposable constellation group (AUDCG). To this end, we first derive a fast noncoherent ML detection algorithm for the received sum constellation and a corresponding closed form symbol error probability formula (SEP). Thus, two important concepts: full receiver diversity and geometrical coding gain, are introduced to quantitatively characterize how quickly SEP decays when the number of base station antennas goes to infinity. Then, the design of an optimal AUDCG is formally formulated into an optimization problem that maximizes both the receiver diversity and the geometrical coding gains subject to each user power constraint. Particularly for a two user system, with each user transmitting nonnegative binary constellation, an optimal closed form sum constellation is obtained. Computer simulations demonstrate that our proposed optimal constellation design substantially outperforms the currently available constellation designs for the same uplink massive MIMO system.","PeriodicalId":152556,"journal":{"name":"2019 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (PACRIM)","volume":"167 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133886143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}