This paper summarizes our recent studies on architecture, photonic integration, system validation and networking performance analysis of a flexible low-latency interconnect optical network switch (Flex-LIONS) for datacenter and high-performance computing (HPC) applications. Flex-LIONS leverages the all-to-all wavelength routing property in arrayed waveguide grating routers (AWGRs) combined with microring resonator (MRR)-based add/drop filtering and multi-wavelength spatial switching to enable topology and bandwidth reconfigurability to adapt the interconnection to different traffic profiles. By exploiting the multiple free spectral ranges of AWGRs, it is also possible to provide reconfiguration while maintaining minimum-diameter all-to-all interconnectivity. We report experimental results on the design, fabrication, and system testing of 8×8 silicon photonic (SiPh) Flex-LIONS chips demonstrating error-free all-to-all communication and reconfiguration exploiting different free spectral ranges (FSR0 and FSR1, respectively). After reconfiguration in FSR1, the bandwidth between the selected pair of nodes is increased from 50Gb/s to 125Gb/s while an all interconnectivity at 25Gb/s is maintained using FSR0. Finally, we investigate the use of Flex-LIONS in two different networking scenarios. First, networking simulations for a 256-node datacenter inter-rack communication scenario show the potential latency and energy benefits when using Flex-LIONS for optical reconfiguration based on different traffic profiles (a legacy fat-tree architecture is used for comparison). Second, we demonstrate the benefits of leveraging two FSRs in an 8-node 64-core computing system to provide reconfiguration for the hotspot nodes while maintaining minimum-diameter all-to-all interconnectivity.

Currently, the top-k error ratio is one of the primary methods to measure the accuracy of multi-category classification. Top-k multiclass SVM was designed to minimize the empirical risk based on the top-k error ratio. Two SDCA-based algorithms exist for learning the top-k SVM, both of which have several desirable properties for achieving optimization. However, both algorithms suffer from a serious disadvantage, that is, they cannot attain the optimal convergence in most cases owing to their theoretical imperfections. As demonstrated through numerical simulations, if the modified SDCA algorithm is employed, optimal convergence is always achieved, in contrast to the failure of the two existing SDCA-based algorithms. Finally, our analytical results are presented to clarify the significance of these existing algorithms.

High dynamic range (HDR) imaging refers to digital image processing that modifies the range of color and contrast to enhance image visibility. To create an HDR image, two or more images that include various information are needed. In order to convert low dynamic range (LDR) images to HDR images, we consider the possibility of using a generative adversarial network (GAN) as an appropriate deep neural network. Deep learning requires a great deal of data in order to build a module, but once the module is created, it is convenient to use. In this paper, we propose a weight map for local luminance based on learning to reconstruct locally tone-mapped images.

Two-ray (TR) fading model is one of the fading models to represent a worst-case fading scenario. We derive the exact closed-form expressions of the generalized moment generating function (G-MGF) for the TR fading model, which enables us to analyze the numerous types of wireless communication applications. Among them, we carry out several analytical results for the TR fading model, including the exact ergodic capacity along with asymptotic expressions and energy detection performance. Finally, we provide numerical results to validate our evaluations.

This paper introduces a novel skin-stretcher device for gently urging head rotation. The device pulls and/or pushes the skin on the user's neck by using servo motors. The user is induced to rotate his/her head based on the sensation caused by the local stretching of skin. This mechanism informs the user when and how much the head rotation is requested; however it does not force head rotation, i.e., it allows the user to ignore the stimuli and to maintain voluntary movements. We implemented a prototype device and analyzed the performance of the skin stretcher as a human-in-the-loop system. Experimental results define its fundamental characteristics, such as input-output gain, settling time, and other dynamic behaviors. Features are analyzed, for example, input-output gain is stable within the same installation condition, but various between users.

Impact of sampling frequency and the number of quantization bit of analog-to-digital conversion (ADC) in a direct detection lightwave system using Kramers-Kronig (KK) relation, which has been attracting attention in recent years, are numerically investigated. We studied the effect of spectral broadening caused by nonlinear operations (logarithm, square root) of the KK algorithm when the frequency gap (shift frequency) between the modulated signal and the optical tone is varied. We found that reception performances depend on both the ADC bandwidth and the relative positions of the optical tone and the spectrum. Spectral broadening caused by the logarithm operation of the KK algorithm is found to be the dominant factor of signal distortion in an ADC bandwidth limited system. We studied the effect of the number of quantization bit on the error vector magnitude (EVM) of KK relation based reception in a carrier-to-signal power ratio (CSPR) adjustable transmission system. We found that performances of KK relation based receiver can be improved by increasing the number of quantization bits. For minimum-phase-condition satisfied KK receiver, the required number of quantization bit was found to be 5 bits or more for detection of QPSK, 16-QAM and 64-QAM-modulated signal after 20-km transmission.

Massive multiple-input multiple-output (MIMO) is an enabling technology for next-generation wireless systems because it provides significant improvements in data rates compared to existing small-scale MIMO systems. However, the increased number of antennas results in high computational complexity for data detection, and requires more efficient detection algorithms. In this paper, we propose a new data detector based on a box-constrained complex-valued dichotomous coordinate descent (BCC-DCD) algorithm for large-scale MIMO systems. The proposed detector involves two steps. First, a transmitted data vector is detected using the BCC-DCD algorithm with a large number of iterations and high solution precision. Second, an improved soft output is generated by reapplying the BCC-DCD algorithm, but with a considerably smaller number of iterations and 1-bit solution precision. Numerical results demonstrate that the proposed method outperforms existing advanced detectors while possessing lower complexity. Specifically, the proposed method provides significantly better detection performance than a BCC-DCD algorithm with similar complexity. The performance advantage increases as the signal-to-noise ratio and the system size increase.

We show that imperfection in an IQ-modulator degrades the accuracy of the auto bias control (ABC) circuit connected to the modulator's complementary port. Theoretical analyses show that the IQ-modulator constructed by a nested Mach-Zehnder modulator with a low extinction ratio can distort a constellation of modulated light observed at the complementary port. We propose an auto calibration technique for the ABC circuit that can effectively suppress this degradation. Experimental results using 32-Gbaud, 16-QAM signals showed the measured Q-factor improved by 0.5dB with our proposed technique.

In this paper, we propose an accurate calibration method for glassless stereoscopic systems. The method uses a lenticular lens on a general display. Glassless stereoscopic displays are currently used in many fields; however, accurately adjusting their physical display position is difficult because an accuracy of several microns or one hundredth of a degree is required, particularly given their larger display area. The proposed method enables a dynamic adjustment of the positions of images on the display to match various physical conditions in three-dimensional (3D) displays. In particular, compared with existing approaches, this avoids degradation of the image quality due to the image location on the screen while improving the image quality by local mapping. Moreover, it is shown to decrease the calibration time by performing simultaneous processing for each local area. As a result of the calibration, the offset jitter representing the crosstalk reduces from 14.946 to 8.645 mm. It is shown that high-quality 3D videos can be generated. Finally, we construct a stereoscopic viewing system using a high-resolution display and lenticular lens and produce high-quality 3D images with automatic calibration.

This paper discusses how to compute the parametric sensitivity function in continuous-time Markov chains (CTMC). The sensitivity function is the first derivative of the steady-state probability vector regarding a CTMC parameter. Since the sensitivity function is given as a solution of linear equations with a sparse matrix, several linear equation solvers are available to obtain it. In this paper, we consider Jacobi and successive-over relaxation as variants of the Gauss-Seidel algorithm. In addition, we develop an algorithm based on the Takahashi method for the sensitivity function. In numerical experiments, we comprehensively evaluate the performance of these algorithms from the viewpoint of computation time and accuracy.

This paper proposes a noise reduction technology for a specific frequency band that uses the pulse coding controlled method to automatically set the notch frequency in DC-DC switching converters of communication equipment. For reducing the power levels at the frequency and its harmonics in the switching converter, we often use a frequency-modulated clock. This paper investigates a technology that prevents modulated clock frequency noise from spreading into protected frequency bands; this proposed noise reduction technology does not distribute the switching noise into some specified frequency bands. The notch in the spectrum of the switching pulses is created by the Pulse Width Coding (PWC) method. In communication devices, the noise in the receiving signal band must be as small as possible. The notch frequency is automatically set to the frequency of the received signal by adjusting the clock frequency using the equation Fn = (P+0.5)Fck. Here Fn is the notch frequency, Fck is the clock frequency, and P is a positive integer that determines the noise spectrum location. Therefore, simply be setting the notch frequency to the received signal frequency can suppress the noise present. We confirm with simulations that the proposed technique is effective for noise reduction and notch generation. Also we implement a method of automatic switching between two receiving channels. The conversion voltage ratio in the pulse width coding method switching converter is analyzed and full automatic notch frequency generation is realized. Experiments on a prototype circuit confirm notch frequency generation.

We have developed an adaptation method which allows the customization of example-based dialog systems for individual users by applying “plus” and “minus” operations to the distributed representations obtained using the word2vec method. After retrieving user-related profile information from the Web, named entity extraction is applied to the retrieval results. Words with a high term frequency-inverse document frequency (TF-IDF) score are then adopted as user related words. Next, we calculate the similarity between the distrubuted representations of selected user-related words and nouns in the existing example phrases, using word2vec embedding. We then generate phrases adapted to the user by substituting user-related words for highly similar words in the original example phrases. Word2vec also has a special property which allows the arithmetic operations “plus” and “minus” to be applied to distributed word representations. By applying these operations to words used in the original phrases, we are able to determine which user-related words can be used to replace the original words. The user-related words are then substituted to create customized example phrases. We evaluated the naturalness of the generated phrases and found that the system could generate natural phrases.

Troika is a recently proposed sponge-based hash function for IOTA's ternary architecture and platform, which is developed by CYBERCRYPT and is now used in IOTA's blockchain. In this paper, we introduce the preimage attack on 2/3 rounds of Troika with a divide-and-conquer approach. Firstly, we propose the equivalent conditions to determine whether a message is the preimage with an algebraic method. As a result, for the preimage attack on two-round Troika, we can search the preimage only in a valid smaller space and efficiently enumerate the messages which can satisfy most of the equivalent conditions with a guess-and-determine technique. Our experiments show that the time complexity of the preimage attack on 2-round Troika can be improved to 379 from 3243. For the preimage attack on 3-round Troika, the MILP-based method is applied to achieve the optimal time complexity, which is 327 times faster than brute force.

The test of homogeneity for normal mixtures has been used in various fields, but its theoretical understanding is limited because the parameter set for the null hypothesis corresponds to singular points in the parameter space. In this paper, we shed a light on this issue from a new perspective, variational Bayes, and offer a theory for testing homogeneity based on it. Conventional theory has not reveal the stochastic behavior of the variational free energy, which is necessary for constructing a hypothesis test, has remained unknown. We clarify it for the first time and construct a new test base on it. Numerical experiments show the validity of our results.

A compact optical polarization converter (PC) based on slot waveguide has been proposed in this study. Utilizing the high refractive index contrast between a Si waveguide and SiO2 cladding on the silicon-on-insulator platform, the light beam can be strongly confined in a slot waveguide structure. The proposed PC consists of a square waveguide and an L-shape cover waveguide. Since the overall structure is symmetrically distributed along the axis rotated 45-degree from the horizontal direction, the optical axis of this PC lies in the direction with equi-angle from two orthogonally polarized modes of the input and output ends, which leads to a high polarization conversion efficiency (PCE). 3D FDTD simulation results illustrate that a TE-to-TM mode conversion is achieved with a device length of 8.2 µm, and the PCE exceeds 99.8%. The structural tolerance and wavelength dependence of the PC have also been discussed in detail.

When using a gesture-based interface for pointing to targets on a wide screen, displaying a large pointer instead of a typical spot pattern reduces disturbance caused by measurement errors of user's pointing posture. However, it remains unclear why a large pointer helps facilitate easy pointing. To examine this issue, in this study we propose a mathematical model that formulates human pointing motions affected by a large pointer. Our idea is to describe the effect of the large pointer as human visual perception, because the user will perceive the pointer-target distance as being shorter than it actually is. We embedded this scheme, referred to as non-linear distance filter (NDF), into a typical feedback loop model designed to formulate human pointing motions. We also proposed a method to estimate NDF mapping from pointing trajectories, and used it to investigate the applicability of the model under three typical disturbance patterns: small vibration, smooth shift, and step signal. Experimental results demonstrated that the proposed NDF-based model could accurately reproduced actual pointing trajectories, achieving high similarity values of 0.89, 0.97, and 0.91 for the three respective disturbance patterns. The results indicate the applicability of the proposed method. In addition, we confirmed that the obtained NDF mappings suggested rationales for why a large pointer helps facilitate easy pointing.

Multi-cycle Test looks promising a way to reduce the test application time of POST (Power-on Self-Test) for achieving a targeted high fault coverage specified by ISO26262 for testing automotive devices. In this paper, we first analyze the mechanism of Stuck-at Fault Detection Degradation problem in multi-cycle test. Based on the result of our analysis we propose a novel solution named FF-Control Point Insertion technique (FF-CPI) to achieve the reduction of scan-in patterns by multi-cycle test. The FF-CPI technique modifies the captured values of scan Flip-Flops (FFs) during capture operation by directly reversing the value of partial FFs or loading random vectors. The FF-CPI technique enhances the number of detectable stuck-at faults under the capture patterns. The experimental results of ISCAS89 and ITC99 benchmarks validated the effectiveness of FF-CPI technique in scan-in pattern reduction for POST.

The communications quality of content delivery networks (CDNs), which are geographically distributed networks that have been optimized for content delivery, deteriorates when interflow congestion conditions are severe. Herein, we propose an adaptive server and path switching scheme that is based on the estimated acquisition throughput of each path. We also provide simulation results that show our proposed method can provide higher throughput performance levels than existing methods.

This paper proposes a voice conversion (VC) method based on a model that links linguistic and acoustic representations via latent phonological distinctive features. Our method, called speech chain VC, is inspired by the concept of the speech chain, where speech communication consists of a chain of events linking the speaker's brain with the listener's brain. We assume that speaker identity information, which appears in the acoustic level, is embedded in two steps — where phonological information is encoded into articulatory movements (linguistic to physiological) and where articulatory movements generate sound waves (physiological to acoustic). Speech chain VC represents these event links by using an adaptive restricted Boltzmann machine (ARBM) introducing phoneme labels and acoustic features as two classes of visible units and latent phonological distinctive features associated with articulatory movements as hidden units. Subjective evaluation experiments showed that intelligibility of the converted speech significantly improved compared with the conventional ARBM-based method. The speaker-identity conversion quality of the proposed method was comparable to that of a Gaussian mixture model (GMM)-based method. Analyses on the representations of the hidden layer of the speech chain VC model supported that some of the hidden units actually correspond to phonological distinctive features. Final part of this paper proposes approaches to achieve one-shot VC by using the speech chain VC model. Subjective evaluation experiments showed that when a target speaker is the same gender as a source speaker, the proposed methods can achieve one-shot VC based on each single source and target speaker's utterance.

For low-density parity-check (LDPC) codes, the penalized decoding method based on the alternating direction method of multipliers (ADMM) can improve the decoding performance at low signal-to-noise ratios and also has low decoding complexity. There are three effective methods that could increase the ADMM penalized decoding speed, which are reducing the number of Euclidean projections in ADMM penalized decoding, designing an effective penalty function and selecting an appropriate layered scheduling strategy for message transmission. In order to further increase the ADMM penalized decoding speed, through reducing the number of Euclidean projections and using the vertical layered scheduling strategy, this paper designs a fast converging ADMM penalized decoding method based on the improved penalty function. Simulation results show that the proposed method not only improves the decoding performance but also reduces the average number of iterations and the average decoding time.