Michael HENTSCHEL Marc DELCROIX Atsunori OGAWA Tomoharu IWATA Tomohiro NAKATANI
Language models are a key technology in various tasks, such as, speech recognition and machine translation. They are usually used on texts covering various domains and as a result domain adaptation has been a long ongoing challenge in language model research. With the rising popularity of neural network based language models, many methods have been proposed in recent years. These methods can be separated into two categories: model based and feature based adaptation methods. Feature based domain adaptation has compared to model based domain adaptation the advantage that it does not require domain labels in the corpus. Most existing feature based adaptation methods are based on bias adaptation. We propose a novel feature based domain adaptation technique using hidden layer factorisation. This method is fundamentally different from existing methods because we use the domain features to calculate a linear combination of linear layers. These linear layers can capture domain specific information and information common to different domains. In the experiments, we compare our proposed method with existing adaptation methods. The compared adaptation techniques are based on two different ideas, that is, bias based adaptation and gating of hidden units. All language models in our comparison use state-of-the-art long short-term memory based recurrent neural networks. We demonstrate the effectiveness of the proposed method with perplexity results for the well-known Penn Treebank and speech recognition results for a corpus of TED talks.
This paper presents a 6th-order quadrature bandpass delta sigma AD modulator (QBPDSM) with 2nd-order image rejection using dynamic amplifier and noise coupling (NC) SAR quantizer embedded by passive adder for the application of wireless communication system. A novel complex integrator using dynamic amplifier is proposed to improve the energy efficiency of the QBPDSM. The NC SAR quantizer can realize an additional 2nd-order noise shaping and 2nd-order image rejection by the digital domain noise coupling technique. As a result, the 6th-order QBPDSM with 2nd-order image rejection is realized by two complex integrators using dynamic amplifier and the NC SAR quantizer. The SPICE simulation results demonstrate the feasibility of the proposed QBPDSM in 90nm CMOS technology. Simulated SNDR of 76.30dB is realized while a sinusoid -3.25dBFS input is sampled at 33.3MS/s and the bandwidth of 2.083MHz (OSR=8) is achieved. The total power consumption in the modulator is 6.74mW while the supply voltage is 1.2V.
Haruya ISHIKAWA Yukitoshi SANADA
This paper evaluates the throughput of a distributed antenna network (DAN) with multiple mobile terminal scheduling and the usage of joint maximum-likelihood detection (MLD). Mobile terminals are closer to the desired antennas in the DAN which leads to higher throughput and better frequency utilization efficiency. However, when multiple mobile terminal scheduling is applied to the DAN, interference can occur between transmitted signals from antennas. Therefore, in this research, mobile terminal scheduling along with joint MLD is applied to reduce the effects of interference. A system level simulation shows that the usage of joint MLD in a densely packed DAN provides better system throughput regardless of the numbers of mobile terminals and fading channels.
Superconducting nanowire single-photon detector(SNSPD) has been one of the important ingredients for photonic quantum information processing (QIP). In order to see the potential of SNSPDs, I briefly review recent progresses of the photonic QIP with SNSPDs implemented for various purposes and present a possible direction for the development of SNSPDs.
Aye Mon HTUN Maung SANN MAW Iwao SASASE
Multiuser massive multi-input multi-output (MU massive MIMO) is considered as a promising technology for the fifth generation (5G) of the wireless communication system. In this paper, we propose a low-complexity joint antenna and user selection scheme with block diagonalization (BD) precoding for MU massive MIMO downlink channel in the time division duplex (TDD) system. The base station (BS) is equipped with a large-scale transmit antenna array while each user is using the single receive antenna in the system. To reduce the hardware cost, BS will be implemented by limited number of radio frequency (RF) chains and BS must activate some selected transmit antennas in the BS side for data transmitting and some users' receive antennas in user side for data receiving. To achieve the reduction in the computation complexity in the antenna and user selection while maintaining the same or higher sum-rate in the system, the proposed scheme relies on three complexity reduction key factors. The first key factor is that finding the average channel gains for the transmit antenna in the BS side and the receive antenna in the user side to select the best channel gain antennas and users. The second key factor called the complexity control factor ξ(Xi) for the antenna set and the user set limitation is used to control the complexity of the brute force search. The third one is that using the assumption of the point-to-point deterministic MIMO channel model to avoid the singular value decomposition (SVD) computation in the brute force search. We show that the proposed scheme offers enormous reduction in the computation complexity while ensuring the acceptable performance in terms of total system sum-rate compared with optimal and other conventional schemes.
Siyu CHEN Ning WANG Mengmeng ZHANG
We propose to discover approximate primary functional dependency (aPFD) for web tables, which focus on the determination relationship between primary attributes and non-primary attributes and are more helpful for entity column detection and topic discovery on web tables. Based on association rules and information theory, we propose metrics Conf and InfoGain to evaluate PFDs. By quantifying PFDs' strength and designing pruning strategies to eliminate false positives, our method could select minimal non-trivial approximate PFD effectively and are scalable to large tables. The comprehensive experimental results on real web datasets show that our method significantly outperforms previous work in both effectiveness and efficiency.
Takahiro HIRAYAMA Masahiro JIBIKI Hiroaki HARAI
Software-defined networking (SDN) technology enables us to flexibly configure switches in a network. Previously, distributed SDN control methods have been discussed to improve their scalability and robustness. Distributed placement of controllers and backing up each other enhance robustness. However, these techniques do not include an emergency measure against large-scale failures such as network separation induced by disasters. In this study, we first propose a network partitioning method to create a robust control plane (C-Plane) against large-scale failures. In our approach, networks are partitioned into multiple sub-networks based on robust topology coefficient (RTC). RTC denotes the probability that nodes in a sub-network isolate from controllers when a large-scale failure occurs. By placing a local controller onto each sub-network, 6%-10% of larger controller-switch connections will be retained after failure as compared to other approaches. Furthermore, we discuss reactive emergency reconstruction of a distributed SDN C-plane. Each node detects a disconnection to its controller. Then, C-plane will be reconstructed by isolated switches and managed by the other substitute controller. Meanwhile, our approach reconstructs C-plane when network connectivity recovers. The main and substitute controllers detect network restoration and merge their C-planes without conflict. Simulation results reveal that our proposed method recovers C-plane logical connectivity with a probability of approximately 90% when failure occurs in 100 node networks. Furthermore, we demonstrate that the convergence time of our reconstruction mechanism is proportional to the network size.
Lina GONG Shujuan JIANG Qiao YU Li JIANG
Heterogeneous defect prediction (HDP) is to detect the largest number of defective software modules in one project by using historical data collected from other projects with different metrics. However, these data can not be directly used because of different metrics set among projects. Meanwhile, software data have more non-defective instances than defective instances which may cause a significant bias towards defective instances. To completely solve these two restrictions, we propose unsupervised deep domain adaptation approach to build a HDP model. Specifically, we firstly map the data of source and target projects into a unified metric representation (UMR). Then, we design a simple neural network (SNN) model to deal with the heterogeneous and class-imbalanced problems in software defect prediction (SDP). In particular, our model introduces the Maximum Mean Discrepancy (MMD) as the distance between the source and target data to reduce the distribution mismatch, and use the cross-entropy loss function as the classification loss. Extensive experiments on 18 public projects from four datasets indicate that the proposed approach can build an effective prediction model for heterogeneous defect prediction (HDP) and outperforms the related competing approaches.
The recent rapid increase in the scale of superconducting quantum computing systems greatly increases the demand for qubit control by digital circuits operating at qubit temperatures. In this paper, superconducting digital circuits, such as single-flux quantum and adiabatic quantum flux parametron circuits are described, that are promising candidates for this purpose. After estimating their energy consumption and speed, a conceptual overview of the superconducting electronics for controlling a multiple-qubit system is provided, as well as some of its component circuits.
Hiroto YASUMI Fukuhito OOSHITA Ken'ichi YAMAGUCHI Michiko INOUE
In this paper, we consider a uniform bipartition problem in a population protocol model. The goal of the uniform bipartition problem is to divide a population into two groups of the same size. We study the problem under global fairness with various assumptions: 1) a population with or without a base station, 2) symmetric or asymmetric protocols, and 3) designated or arbitrary initial states. As a result, we completely clarify solvability of the uniform bipartition problem under global fairness and, if solvable, show the tight upper and lower bounds on the number of states.
Akihiro MATSUURA Yoshiaki SHOJI
In this paper, we show the explicit formula of the recurrence relation for the Tower of Hanoi on the star graph with four vertices, where the perfect tower of disks on a leaf vertex is transferred to the central vertex. This gives the solution to the problem posed at the 17th International Conference on Fibonacci Numbers and Their Applications[11]. Then, the recurrence relation are generalized to include the ones for the original 4-peg Tower of Hanoi and the Star Tower of Hanoi of transferring the tower from a leaf to another.
Xiaoqiong ZHAO Shanping LI Huan YU Ye WANG Weiwei QIU
Background: The applying of third-party libraries is an integral part of many applications. But the libraries choosing is time-consuming even for experienced developers. The automated recommendation system for libraries recommendation is widely researched to help developers to choose libraries. Aim: from software engineering aspect, our research aims to give developers a reliable recommended list of third-party libraries at the early phase of software development lifecycle to help them build their development environment faster; and from technical aspect, our research aims to build a generalizable recommendation system framework which combines collaborative filtering and topic modeling techniques, in order to improve the performance of libraries recommendation significantly. Our works on this research: 1) we design a hybrid methodology to combine collaborative filtering and LDA text mining technology; 2) we build a recommendation system framework successfully based on the above hybrid methodology; 3) we make a well-designed experiment to validate the methodology and framework which use the data of 1,013 mobile application projects; 4) we do the evaluation for the result of the experiment. Conclusions: 1) hybrid methodology with collaborative filtering and LDA can improve the performance of libraries recommendation significantly; 2) based on the hybrid methodology, the framework works very well on the libraries recommendation for helping developers' libraries choosing. Further research is necessary to improve the performance of the libraries recommendation including: 1) use more accurate NLP technologies improve the correlation analysis; 2) try other similarity calculation methodology for collaborative filtering to rise the accuracy; 3) on this research, we just bring the time-series approach to the framework and make an experiment as comparative trial, the result shows that the performance improves continuously, so in further research we plan to use time-series data-mining as the basic methodology to update the framework.
Ying CAO Lijuan SUN Chong HAN Jian GUO
Due to the inevitable data missing problem during visual data acquisition, the recovery of color images and videos from limited useful information has become an important topic, for which tensor completion has been proved to be a promising solution in previous studies. In this paper, we propose a novel completion scheme, which can effectively recover missing entries in color images and videos represented by tensors. We first employ a modified tensor train (TT) decomposition as tensor approximation scheme in the concept of TT rank to generate better-constructed and more balanced tensors which preserve only relatively significant informative data in tensors of visual data. Afterwards, we further introduce a TT rank-based weight scheme which can define the value of weights adaptively in tensor completion problem. Finally, we combine the two schemes with Simple Low Rank Tensor Completion via Tensor Train (SiLRTC-TT) to construct our completion algorithm, Low Rank Approximated Tensor Completion via Adaptive Tensor Train (LRATC-ATT). Experimental results validate that the proposed approach outperforms typical tensor completion algorithms in recovering tensors of visual data even with high missing ratios.
Fei WU Xiwei DONG Lu HAN Xiao-Yuan JING Yi-mu JI
Recently, multi-view dictionary learning technique has attracted lots of research interest. Although several multi-view dictionary learning methods have been addressed, they can be further improved. Most of existing multi-view dictionary learning methods adopt the l0 or l1-norm sparsity constraint on the representation coefficients, which makes the training and testing phases time-consuming. In this paper, we propose a novel multi-view dictionary learning approach named multi-view synthesis and analysis dictionaries learning (MSADL), which jointly learns multiple discriminant dictionary pairs with each corresponding to one view and containing a structured synthesis dictionary and a structured analysis dictionary. MSADL utilizes synthesis dictionaries to achieve class-specific reconstruction and uses analysis dictionaries to generate discriminative code coefficients by linear projection. Furthermore, we design an uncorrelation term for multi-view dictionary learning, such that the redundancy among synthesis dictionaries learned from different views can be reduced. Two widely used datasets are employed as test data. Experimental results demonstrate the efficiency and effectiveness of the proposed approach.
Yibo JIANG Hui BI Hui LI Zhihao XU
The 3D measurement is widely required in modern industries. In this letter, a method based on the RGBD saliency detection with depth range adjusting (RGBD-DRA) is proposed for 3D measurement. By using superpixels and prior maps, RGBD saliency detection is utilized to detect and measure the target object automatically Meanwhile, the proposed depth range adjusting is processing while measuring to prompt the measuring accuracy further. The experimental results demonstrate the proposed method automatic and accurate, with 3 mm and 3.77% maximum deviation value and rate, respectively.
Kunho PARK Min Joo JEONG Jong Jin BAEK Se Woong KIM Youn Tae KIM
This paper presents the bit error rate (BER) performance of human body communication (HBC) receivers in interference-rich environments. The BER performance was measured while applying an interference signal to the HBC receiver to consider the effect of receiver performance on BER performance. During the measurement, a signal attenuator was used to mimic the signal loss of the human body channel, which improved the repeatability of the measurement results. The measurement results showed that HBC is robust against the interference when frequency selective digital transmission (FSDT) is used as a modulation scheme. The BER performance in this paper can be effectively used to evaluate a communication performance of HBC.
Tongxin YANG Tomoaki UKEZONO Toshinori SATO
Many applications, such as image signal processing, has an inherent tolerance for insignificant inaccuracies. Multiplication is a key arithmetic function for many applications. Approximate multipliers are considered an efficient technique to trade off energy relative to performance and accuracy for the error-tolerant applications. Here, we design and analyze four approximate multipliers that demonstrate lower power consumption and shorter critical path delay than the conventional multiplier. They employ an approximate tree compressor that halves the height of the partial product tree and generates a vector to compensate accuracy. Compared with the conventional Wallace tree multiplier, one of the evaluated 8-bit approximate multipliers reduces power consumption and critical path delay by 36.9% and 38.9%, respectively. With a 0.25% normalized mean error distance, the silicon area required to implement the multiplier is reduced by 50.3%. Our multipliers outperform the previously proposed approximate multipliers relative to power consumption, critical path delay, and design area. Results from two image processing applications also demonstrate that the qualities of the images processed by our multipliers are sufficiently accurate for such error-tolerant applications.
Zuohong XU Jiang ZHU Qian CHENG Zixuan ZHANG
Quasi cyclic LDPC (QC-LDPC) codes consisting of circulant permutation matrices (CPM-QC-LDPC) are one of the most attractive types of LDPC codes due to their many advantages. In this paper, we mainly do some research on CPM-QC-LDPC codes. We first propose a two-stage decoding scheme mainly based on parity check matrix transform (MT), which can efficiently improve the bit error rate performance. To optimize the tradeoff between hardware implementation complexity and decoding performance, an improved method that combines our proposed MT scheme with the existing CPM-RID decoding scheme is presented. An experiment shows that both schemes can improve the bit error rate (BER) performance. Finally, we show that the MT decoding mechanism can be applied to other types of LDPC codes. We apply the MT scheme to random LDPC codes and show that it can efficiently lower the error floor.
Xiaoxia LIU Degen HUANG Zhangzhi YIN Fuji REN
Collocation is a ubiquitous phenomenon in languages and accurate collocation recognition and extraction is of great significance to many natural language processing tasks. Collocations can be differentiated from simple bigram collocations to collocation frames (referring to distant multi-gram collocations). So far little focus is put on collocation frames. Oriented to translation and parsing, this study aims to recognize and extract the longest possible collocation frames from given sentences. We first extract bigram collocations with distributional semantics based method by introducing collocation patterns and integrating some state-of-the-art association measures. Based on bigram collocations extracted by the proposed method, we get the longest collocation frames according to recursive nature and linguistic rules of collocations. Compared with the baseline systems, the proposed method performs significantly better in bigram collocation extraction both in precision and recall. And in extracting collocation frames, the proposed method performs even better with the precision similar to its bigram collocation extraction results.
Takayoshi HIRASAWA Shigeyuki AKIBA Jiro HIROKAWA Makoto ANDO
This paper studies the performance of the quantitative RF power variation in Radio-over-Fiber beam forming system utilizing a phased array-antenna integrating photo-diodes in downlink network for next generation millimeter wave band radio access. Firstly, we described details of fabrication of an integrated photonic array-antenna (IPA), where a 60GHz patch antenna 4×2 array and high-speed photo-diodes were integrated into a substrate. We evaluated RF transmission efficiency as an IPA system for Radio-over-Fiber (RoF)-based mobile front hall architecture with remote antenna beam forming capability. We clarified the characteristics of discrete and integrated devices such as an intensity modulator (IM), an optical fiber and the IPA and calculated RF power radiated from the IPA taking account of the measured data of the devices. Based on the experimental results on RF tone signal transmission by utilizing the IPA, attainable transmission distance of wireless communication by improvement and optimization of the used devices was discussed. We deduced that the antenna could output sufficient power when we consider that the cell size of the future mobile communication systems would be around 100 meters or smaller.