Ordinal classification is a class of special tasks in machine learning and pattern recognition. As to ordinal classification, there is an ordinal structure among different decision values. The monotonicity constraint between features and decision should be taken into account as the fundamental assumption. However, in real-world applications, this assumption may be not true. Only some candidate features, instead of all, are monotonic with decision. So the existing feature selection algorithms which are designed for nominal classification or monotonic classification are not suitable for ordinal classification. In this paper, we propose a feature selection algorithm for ordinal classification based on considering the non-monotonic and monotonic features separately. We first introduce an assumption of hybrid monotonic classification consistency and define a feature evaluation function to calculate the relevance between the features and decision for ordinal classification. Then, we combine the reported measure and genetic algorithm (GA) to search the optimal feature subset. A collection of numerical experiments are implemented to show that the proposed approach can effectively reduce the feature size and improve the classification performance.

This paper presents the optimal implementation methods for 256-bit elliptic curve digital signature algorithm (ECDSA) signature generation processors with high speed Montgomery multipliers. We have explored the radix of the data path of the Montgomery multiplier from 2-bit to 256-bit operation and proposed the use of pipelined Montgomery multipliers for signature generation speed, area, and energy optimization. The key factor in the design optimization is how to perform modular multiplication. The high radix Montgomery multiplier is known to be an efficient implementation for high-speed modular multiplication. We have implemented ECDSA signature generation processors with high radix Montgomery multipliers using 65-nm SOTB CMOS technology. Post-layout results show that the fastest ECDSA signature generation time of 63.5µs with radix-256-bit, a two-module four-streams pipeline architecture, and an area of 0.365mm2 (which is the smallest) with a radix-16-bit zero-pipeline architecture, and the smallest signature generation energy of 9.51µJ with radix-256-bit zero-pipeline architecture.

This paper presents a wide-input-voltage-range and high-efficiency boost converter that is assisted by a transformer-based oscillator. The oscillator can provide a sufficient amount of power to drive a following switched-inductor boost converter at low voltages. Moreover, it adopts a novel amplitude-regulation circuit (ARC) without using high power-consuming protective devices to suppress the expansion of the oscillation amplitude at high input voltages. Therefore, it can avoid over-voltage problems without sacrificing the power efficiency. Additionally, a power-down circuit (PDC) is implemented to turn off the oscillator, when the boost converter can be driven by its own output power, thus, eliminating the power consumption by the oscillator and improving the power efficiency. We implemented the ARC and the PDC with discrete components rather than one-chip integration for the proof of concept. The experimental results showed that the proposed circuit became possible to operate from an input voltage of 60mV to 3V while maintaining high peak efficiency up to 92%. To the best of our knowledge, this converter provides a wider input range in comparison with the previously-published converters. We are convinced that the proposed approach by inserting an appropriate start-up circuit in a commercial converter will be effective for rapid design proposals in order to respond promptly to customer needs as Internet of things (IoT) devices with energy harvester.

This paper proposes a low power single-ended successive approximation register (SAR) analog-to-digital converter (ADC) to replace the only analog active circuit, the comparator, with a digital circuit, which is an inverter-based comparator. The replacement helps possible design automation. The inverter threshold voltage variation impact is minimal because an SAR ADC has only one comparator, and many applications are either insensitive to the resulting ADC offset or easily corrected digitally. The proposed resetting approach mitigates leakage when the input is close to the threshold voltage. As an intrinsic headroom-free, and thus low-rail-voltage, friendly structure, an inverter-based comparator also occupies a small area. Furthermore, an 11-bit ADC was designed and manufactured through a 0.35-µm CMOS process by adopting a low-power switching procedure. The ADC achieves an FOM of 181fJ/Conv.-step at a 25kS/s sampling rate when the supply voltage VDD is 1.2V.

Patarin proposed a crytographic trapdoor called Hidden Field Equation (HFE), a trapdoor based on the Multivariate Quadratic (MQ) and the Isomorphism of Polynomials (IP) problems. The MQ problem was proved by Patarin et al.'s to be NP-complete. Although the basic HFE has been proved to be vulnerable to attacks, its variants obtained by some modifications have been proved to be stronger against attacks. The Quartz digital signature scheme based on the HFEv- trapdoor (a variant of HFE) with particular choices of parameters, has been shown to be stronger against algebraic attacks to recover the private key. Furthermore, it generates reasonably short signatures. However, Joux et al. proved (based on the Birthday Paradox Attack) that Quartz is malleable in the sense that, if an adversary gets a valid pair of message and signature, a valid signature to another related message is obtainable with 250 computations and 250 queries to the signing oracle. Currently, the recommended minimum security level is 2112. Our signature scheme is also based on Quartz but we achieve a 2112 security level against Joux et al.'s attack. It is also more efficient in signature verification and vector initializations. Furthermore, we implemented both the original and our improved Quartz signature and run empirical comparisons.

The three dimensional (3D) reconstruction of a medical image sequence can provide intuitive morphologies of a target and help doctors to make more reliable diagnosis and give a proper treatment plan. This paper aims to reconstruct the surface of a renal corpuscle from the microscope renal biopsy image sequence. First, the contours of renal corpuscle in all slices are extracted automatically by using a context-based segmentation method with a coarse registration. Then, a new coevolutionary-based strategy is proposed to realize a fine registration. Finally, a Gauss-Seidel iteration method is introduced to achieve a non-rigid registration. Benefiting from the registrations, a smooth surface of the target can be reconstructed easily. Experimental results prove that the proposed method can effectively register the contours and give an acceptable surface for medical doctors.

As a research project supported jointly by the National Institute of Information and Communications Technology (NICT) in Japan and the European Commission under its 7th Framework Program, the GreenICN Project has been in operation from 2013 to 2016. The GreenICN project focused on two typical application scenarios, one a disaster scenario and the other a video delivery scenario. The disaster scenario assumed a situation of limited resources, and the video delivery scenario assumed a situation of large-scale content delivery. In both situations, the project challenged to provide “green”, i.e. energy-efficient, content delivery mechanism. For this goal, we designed an energy consumption model to lay out energy reduction policies. For the achievement of the policies, we improved ICN architecture, for example a name-based publish/subscribe mechanism, an effective cache management policy,energy-efficient security scheme and a new energy API. This paper provides a summary of our achievements and descriptions of some outcome.

Information-centric networking (ICN) has been investigated as a new communication model that is optimal for data registration and retrieval. A promising application of ICN is mobile machine-to-machine (M2M) communication in which data are registered by M2M terminals, such as vehicles, and retrieved by other M2M terminals. One of the most difficult challenges with ICN is achieving data mobility in which the data are registered by moving terminals and the location of the data changes constantly. To gain access to moving data, the data retrieval messages must access the routing information, which results in a high volume of message transaction loads of high-tier nodes such as the name resolution nodes. We previously proposed a scheme called data-centric network (DCN), which mitigates this problem by allocating multiple intermediate nodes that act as route aggregation points and by establishing optimized routes. In this paper, we compare the transaction load of DCN with those of conventional ICN schemes using theoretical evaluation based on probability calculation. We also compare the amount of route information and transaction loads using a simulator against binary tree and ISP backbone topologies. From these evaluations, we clarify the characteristics of each ICN scheme in different terminal distribution and communication patterns and show that DCN reduces the transaction loads of high-tier nodes when the terminals are communicating locally.

We have derived the physics-based equivalent circuit model of a semiconductor-integrated bow-tie antenna (BTA) for expressing its impedance and radiation characteristics as a terahertz transmitter. The equivalent circuit branches and components, consisting of 16 RLC parameters are determined based on electromagnetic simulations. All the values of the circuit elements are identified using the particle swarm optimization (PSO) that is one of the modern multi-purpose optimization methods. Moreover, each element value can also be explained by the structure of the semiconductor-integrated BTA, the device size, and the material parameters.

Content-centric networking (CCN) is an emerging networking architecture that is being actively investigated in both the research and industrial communities. In the latest version of CCN, a large number of interests have to be issued when large content is retrieved. Since CCN routers have to search several tables for each incoming interest, this could cause a serious problem of router workload. In order to solve this problem, this paper introduces a novel strategy of “grouping” multiple interests with common information and “packing” them to a special interest called the list interest. Our list interest is designed to co-operate with the manifest of CCN as its dual. This paper demonstrates that by skipping and terminating several search steps using the common information in the list interest, the router can search its tables for the list interest-based request with dramatically smaller complexity than the case of the standard interest-based request. Furthermore, we also consider the deployment of list interests and design a novel TCP-like congestion control method for list interests to employ them just like standard interests.

Information-Centric Networking (ICN) technology has recently been attracting substantial interest in the research community as one of the most promising future Internet architectures. The Named Data Networking (NDN) approach, which is one of the most recent instantiations of the ICN approach, would be a good choice for multimedia services, because NDN utilizes in-network storage embedded in NDN routers by caching recently or frequently requested contents. It is important to determine which data to cache at which NDN routers in order to achieve high performance, by considering not only the popularity of contents but also the inter-chunk popularity of a content item. This paper presents a chunk-block-based incremental caching scheme that considers both content and inter-chunk popularity. Our proposed scheme employs an incremental cache populating mechanism, which utilizes not only core-side but also edge-side NDN routers according to the request rate of the content item. Through simulations, we show that the proposed scheme achieves less delay, reduced redundant network traffic, and a higher cache hit ratio than legacy schemes.

In this work, we propose general, practical and accurate models to analyze the performance of multi-cache systems, in which a cache forwards its miss stream (i.e., requests which have not found the target item) to other caches. We extend a miss stream modeling technique originally known as Melazzi's approximation, which provides a simple but accurate approximate analysis for caches with cascade configurations. We consider several practical replication strategies, which have been commonly adopted in the context of ICN, taking into account the effects of temporal locality. Also, we capture the existing state correlations between neighboring caches by exploiting the cache eviction time. Our proposed models to handle traffic patterns allow us to go beyond the standard Poisson approximation under Independent Reference Model. Our results, validated against simulations, provide interesting insights into the performance of multi-cache systems with different replication strategies.

In this paper, we investigate the degrees of freedom (DoF) of a MIMO cellular interfering network (CIN) with L (L≥3) cells and K users per cell. Previous works established the DoF upper bound of LK(M+N)/(LK+1) for the MIMO CIN by analyzing the interference alignment (IA) feasibility, where M and N denote the number of antennas at each base station (BS) and each user, respectively. However, there is still a gap between the DoF upper bound and the achievable DoF in existing designs. To address this problem, we propose two linear IA schemes without symbol extensions to jointly design transmit and receive beamforming matrices to align and eliminate interference. In the two schemes, the transmit beamforming vectors are allocated to different cluster structures so that the inter-cell interference (ICI) data streams from different ICI channels are aligned. The first scheme, named fixed cluster structure (FCS-IA) scheme, allocates ICI beamforming vectors to the cluster structures of fixed dimension and can achieve the DoF upper bound under some system configurations. The second scheme, named dynamic cluster structure IA (DCS-IA) scheme, allocates ICI beamforming vectors to the cluster structures of dynamic dimension and can get a tradeoff between the number of antennas at BSs and users so that ICI alignment can be applied under various system configurations. Through theoretical analysis and numerical simulations, we verify that the DoF upper bound can be achieved by using the FCS-IA scheme. Furthermore, we show that the proposed schemes can provide significant performance gain over the time division multiple access (TDMA) scheme in terms of DoF. From the perspective of DoF, it is shown that the proposed schemes are more effective than the conventional IA schemes for the MIMO CIN.

In this paper, a novel slow-wave half mode substrate integrated waveguide (SW-HMSIW) structure is presented and experimentally demonstrated, and some interesting slow-wave propagation effects are obtained. The SW-HMSIW enables the cutoff frequency reduction and phase velocity to decrease without sacrificing its performance at the same lateral dimension, which equivalently reduces the lateral dimension and longitudinal size at the same frequency. Specifically, with the different loading microstrip width, a cutoff frequency reduction of 16%, 25%, 30% is achieved compared to the conventional HMSIW at the same lateral dimension. Both lateral and longitudinal size reductions significantly extend the operating range of SIW structures to low frequency region.

In this paper, we discuss an algebraic decoding of BCH codes over symbol-pair read channels. The channels output overlapping pairs of symbols in storage applications. The pair distance and pair error are used in the channels. We define a polynomial that represents the positions of the pair errors as the error-locator polynomials and a polynomial that represents the positions of the pairs of a received pair vector in conflict as conflict-locator polynomial. In this paper, we propose algebraic methods for correcting two-pair and three-pair errors for BCH codes. First, we show the relation between the error-locator polynomials and the conflict-locator polynomial. Second, we show the relation among these polynomials and the syndromes. Finally, we provide how to correct the pair errors by solving equations including the relational expression by algebraic methods.

Light-field image processing has been widely employed in many areas, from mobile devices to manufacturing applications. The fundamental process to extract the usable information requires significant computation with high-resolution raw image data. A graphics processing unit (GPU) is used to exploit the data parallelism as in general image processing applications. However, the sparse memory access pattern of the applications reduced the performance of GPU devices for both systematic and algorithmic reasons. Thus, we propose an optimization technique which redesigns the memory access pattern of the applications to alleviate the memory bottleneck of rendering application and to increase the data reusability for depth extraction application. We evaluated our optimized implementations with the state-of-the-art algorithm implementations on several GPUs where all implementations were optimally configured for each specific device. Our proposed optimization increased the performance of rendering application on GTX-780 GPU by 30% and depth extraction application on GTX-780 and GTX-980 GPUs by 82% and 18%, respectively, compared with the original implementations.

File system workloads are increasing write-heavy. The growing capacity of RAM in modern nodes allows many reads to be satisfied from memory while writes must be persisted to disk. Today's sophisticated local file systems like Ext4, XFS and Btrfs optimize for reads but suffer from workloads dominated by microdata (including metadata and tiny files). In this paper we present an LSM-tree-based file system, RFS, which aims to take advantages of the write optimization of LSM-tree to provide enhanced microdata performance, while offering matching performance for large files. RFS incrementally partitions the namespace into several metadata columns on a per-directory basis, preserving disk locality for directories and reducing the write amplification of LSM-trees. A write-ordered log-structured layout is used to store small files efficiently, rather than embedding the contents of small files into inodes. We also propose an optimization of global bloom filters for efficient point lookups. Experiments show our library version of RFS can handle microwrite-intensive workloads 2-10 times faster than existing solutions such as Ext4, Btrfs and XFS.

In this paper, a discrete particle swarm optimization method is proposed to solve the multi-objective task assignment problem in distributed environment. The objectives of optimization include the makespan for task execution and the budget caused by resource occupation. A two-stage approach is designed as follows. In the first stage, several artificial particles are added into the initialized swarm to guide the search direction. In the second stage, we redefine the operators of the discrete PSO to implement addition, subtraction and multiplication. Besides, a fuzzy-cost-based elite selection is used to improve the computational efficiency. Evaluation shows that the proposed algorithm achieves Pareto improvement in comparison to the state-of-the-art algorithms.

Inter-person occlusion handling is a critical issue in the field of tracking, and it has been extensively researched. Several state-of-the-art methods have been proposed, such as focusing on the appearance of the targets or utilizing knowledge of the scene. In contrast with the approaches proposed in the literature, we propose to address this issue using a social interaction model, which allows us to explore spatio-temporal information pertaining to the targets involved in the occlusion situation. Our experimental results show promising results compared with those obtained using other methods.