We propose a domain specific model for statistical machine translation. It is well-known that domain specific language models perform well in automatic speech recognition. We show that domain specific language and translation models also benefit statistical machine translation. However, there are two problems with using domain specific models. The first is the data sparseness problem. We employ an adaptation technique to overcome this problem. The second issue is domain prediction. In order to perform adaptation, the domain must be provided, however in many cases, the domain is not known or changes dynamically. For these cases, not only the translation target sentence but also the domain must be predicted. This paper focuses on the domain prediction problem for statistical machine translation. In the proposed method, a bilingual training corpus, is automatically clustered into sub-corpora. Each sub-corpus is deemed to be a domain. The domain of a source sentence is predicted by using its similarity to the sub-corpora. The predicted domain (sub-corpus) specific language and translation models are then used for the translation decoding. This approach gave an improvement of 2.7 in BLEU score on the IWSLT05 Japanese to English evaluation corpus (improving the score from 52.4 to 55.1). This is a substantial gain and indicates the validity of the proposed bilingual cluster based models.

Convolutional compactors offer a promising technique of compacting test responses. In this study we expand the architecture of convolutional compactor onto a Galois field in order to improve compaction ratio as well as reduce X-masking probability, namely, the probability that an error is masked by unknown values. While each scan chain is independently connected by EOR gates in the conventional arrangement, the proposed scheme treats q signals as an element over GF(2q), and the connections are configured on the same field. We show the arrangement of the proposed compactors and the equivalent expression over GF(2). We then evaluate the effectiveness of the proposed expansion in terms of X-masking probability by simulations with uniform distribution of X-values, as well as reduction of hardware overheads. Furthermore, we evaluate a multi-weight arrangement of the proposed compactors for non-uniform X distributions.

The research on displacement vector detection has gained increasing attention in recent years. However, no relationship between displacement vectors and the outlines of objects in motion has been established. We describe a new method of detecting displacement vectors through edge segment detection by emphasizing the correlation between displacement vectors and their outlines. Specifically, after detecting an edge segment, the direction of motion of the edge segment can be inferred through the variation in the values of the Laplacian-Gaussian filter at the position near the edge segment before and after the motion. Then, by observing the degrees of displacement before and after the motion, the displacement vector can be calculated. The accuracy compared to other methods of displacement vector detection demonstrates the feasibility of this method.

Permutation ambiguity of the classical Independent Component Analysis (ICA) may cause problems in feature extraction for pattern classification. Especially when only a small subset of components is derived from data, these components may not be most distinctive for classification, because ICA is an unsupervised method. We include a selective prior for de-mixing coefficients into the classical ICA to alleviate the problem. Since the prior is constructed upon the classification information from the training data, we refer to the proposed ICA model with a selective prior as a supervised ICA (sICA). We formulated the learning rule for sICA by taking a Maximum a Posteriori (MAP) scheme and further derived a fixed point algorithm for learning the de-mixing matrix. We investigate the performance of sICA in facial expression recognition from the aspects of both correct rate of recognition and robustness even with few independent components.

The Earth will require sustainable electricity sources equivalent to 3 to 5 times the commercial power presently produced by 2050. Solar Power Satellite (SPS) is one option for meeting the huge future energy demand. SPS can send enormous amounts of power to the Earth as the form of microwave (MW). A highly efficient microwave power transmission (MPT) system is needed for SPS. A critical goal of SPS is to maintain highest Beam Efficiency (BE) because the microwaves from SPS will be converted to utility power unlike the MW from communication satellites. Another critical goal of SPS is to maintain Side Lobe Levels (SLL) as small as possible to reduce interference to other communication systems. One way to decrease SLL and increase BE is the edge tapering of a phased array antenna. However, tapering the excitation requires a technically complicated system. Another way of achieving minimum SLL is with randomly spaced element position but it does not guarantee higher BE and the determination of random element position is also a difficult task. Isosceles Trapezoidal Distribution (ITD) edge tapered antenna was studied for SPS as an optimization between full edge tapering and uniform amplitude distribution. The highest Beam Collection Efficiency (BCE) and lowest SLL (except maximum SLL) are possible to achieve in ITD edge tapering and ITD edge tapered antenna is technically better. The performance of ITD is further improved from the perspective of both Maximum Side Lobe Level (MSLL) and BE by using unequal spacing of the antenna elements. A remarkable reduction in MSLL is achieved with ITD edge tapering with Unequal element spacing (ITDU). BE was also highest in ITDU. Determination of unequal element position for ITDU is very easy. ITDU is a newer concept that is experimented for the first time. The merits of ITDU over ITD and Gaussian edge tapering are discussed.

Inspired by the success of the low-density parity-check (LDPC) codes in the field of error-control coding, in this paper we propose transforming the downlink multiuser multiple-input multiple-output scheduling problem into an LDPC-like problem using the normal graph. Based on the normal graph framework, soft information, which indicates the probability that each user will be scheduled to transmit packets at the access point through a specified angle-frequency sub-channel, is exchanged among the local processors to iteratively optimize the multiuser transmission schedule. Computer simulations show that the proposed algorithm can efficiently schedule simultaneous multiuser transmission which then increases the overall channel utilization and reduces the average packet delay.

The average bit error rate performances of M-ary orthogonal code shift keying (CSK) in Rician fading environments are analyzed in this letter. CSK is a digital modulation scheme that uses a code set as M-ary signals. In CSK, one code is selected from a code set containing M codes according to the information data. A signal is modulated by using this code and the effect of fading can be reduced by applying interleaving to the elements of the codes. In the analysis, the bit error probability is derived in closed form expression by using the Chernoff bound. The analysis results show that the error probability decreases when the code length is increased and that an arbitrarily small error probability is achieved as the code length approaches infinity, provided that Eb/N0 exceeds 1.42 dB.

Recently, there has been extensive research on resource allocation schemes for multicast services that would satisfy the requirements of multimedia traffic. Although several schemes have been proposed to improve the performance of individual multicast groups, it is not easy to achieve both throughput efficiency and user fairness. In this study, we propose a new multicast scheduling scheme for achieving proportional fair (PF) allocation in wireless cellular systems. The basic idea of PF is to schedule the user whose corresponding instantaneous channel quality is the highest relative to the average channel condition over a given time scale. We first extend the PF metric to the extent that the scheduler can reflect the user's varying channel gain, and fairness, not only in the unicast case, but also in multicast transmissions. A multicast PF scheme maximizes the summation of the logarithmic average rate of all multicasting users. Thus, it improves the fairness to mobile users when compared to max-rate allocation, because the logarithmic rate gives more weight to lower rate users, while achieving high throughput. Moreover, the proposed scheme is less complex than max-rate allocation.

In this letter, a computational approach for solving cross-coupled algebraic Riccati equations (CAREs) is investigated. The main purpose of this letter is to propose a new algorithm that combines Newton's method with a gradient-based iterative (GI) algorithm for solving CAREs. In particular, it is noteworthy that both a quadratic convergence under an appropriate initial condition and reduction in dimensions for matrix computation are both achieved. A numerical example is provided to demonstrate the efficiency of this proposed algorithm.

This paper presents a methodology for performing on-line voltage risk identification (VRI) in power supply networks using hyperrectangular composite neural networks (HRCNNs) and synchronized phasor measurements. The FHRCNN presented in this study integrates the paradigm of neural networks with the concept of knowledge-based approaches, rendering them both more useful than when applied alone. The fuzzy rules extracted from the dynamic data relating to the power system formalize the knowledge applied by experts when conducting the voltage risk assessment procedure. The efficiency of the proposed technique is demonstrated via its application to the Taiwan Power Provider System (Tai-Power System) under various operating conditions. Overall, the results indicated that the proposed scheme achieves a minimum 97 % success rate in determining the current voltage security level.

We propose and demonstrate a simple one-bus two-ring configuration where the two rings are mutually coupled that has advantages over the one-ring structure. Unlike a one cavity system, it can exhibit near critically-coupled transmission with a broader range of loss. It can also significantly enhance the cavity finesse by simply making the second ring twice the size of the bus-coupled one, with the enhancement proportional to the intensity buildup in the second ring.

Combination of mutually complementary features is necessary to cope with various changes in pattern classification between normal and pathological voices. This paper proposes a method to improve pathological/normal voice classification performance by combining heterogeneous features. Different combinations of auditory-based and higher-order features are investigated. Their performances are measured by Gaussian mixture models (GMMs), linear discriminant analysis (LDA), and a classification and regression tree (CART) method. The proposed classification method by using the CART analysis is shown to be an effective method for pathological voice detection, with a 92.7% classification performance rate. This is a noticeable improvement of 54.32% compared to the MFCC-based GMM algorithm in terms of error reduction.

In this paper, we study a problem of inferring blood relationships which satisfy a given matrix of genetic distances between all pairs of n nodes. Blood relationships are represented by our proposed graph class, which is called a pedigree graph. A pedigree graph is a directed acyclic graph in which the maximum indegree is at most two. We show that the number of pedigree graphs which satisfy the condition of given genetic distances may be exponential, but they can be represented by one directed acyclic graph with n nodes. Moreover, an O(n3) time algorithm which solves the problem is also given. Although phylogenetic trees and phylogenetic networks are similar data structures to pedigree graphs, it seems that inferring methods for phylogenetic trees and networks cannot be applied to infer pedigree graphs since nodes of phylogenetic trees and networks represent species whereas nodes of pedigree graphs represent individuals. We also show an O(n2) time algorithm which detects a contradiction between a given pedigree graph and distance matrix of genetic distances.

Though machine vision systems for automatically detecting visual defects, called mura, have been developed for thin flat transistor liquid crystal display (TFT-LCD) panels, they have not yet reached a level of reliability which can replace human inspectors. To establish an objective criterion for identifying real defects, some index functions for quantifying defect levels based on human perception have been recently researched. However, while these functions have been verified in the laboratory, further consideration is needed in order to apply them to real systems in the field. To begin with, we should correct the distortion occurring through the capturing of panels. Distortion can cause the defect level in the observed image to differ from that in the panel. There are several known methods to restore the observed image in general vision systems. However, TFT-LCD panel images have a unique background degradation composed of background non-uniformity and vignetting effect which cannot easily be restored through traditional methods. Therefore, in this paper we present a new method to correct background degradation of TFT-LCD panel images using principal component analysis (PCA). Experimental results show that our method properly restores the given observed images and the transformed shape of muras closely approaches the original undistorted shape.

For real time image processing, a readout circuit for an infrared focal plane array (IRFPA) involving a new edge detection technique has been proposed in this letter. A non-uniformity correction unit (NUC), essential in an IRFPA because of bad non-uniformity characteristics of IR sensors is eliminated in this circuit by using a noise tolerant edge detection technique. In addition, real time edge detection can be possible, because of pixel-level integration and parallel processing. The proposed readout circuit shows an approximately three to nine times better edge error rate than other available methods using pixel-level parallel processing.

Consider an undirected multigraph G=(V,E) with n vertices and m edges, and let Ni denote the number of connected spanning subgraphs with i(min) edges in G. Recently, we showed in [3] the validity of (m-i+1)Ni-1>Ni for a simple graph and each i(min). Note that, from this inequality, 2 is easily derived. In this paper, for a multigraph G and all i(min), we prove (m-i+1)Ni-1(i-n+2)Ni, and give a necessary and sufficient condition by which (m-i+1)Ni-1=(i-n+2)Ni. In particular, this means that (m-i+1)Ni-1>Ni is not valid for all multigraphs, in general. Furthermore, we prove 2, which is not straightforwardly derived from (m-i+1)Ni-1(i-n+2)Ni, and also introduce a necessary and sufficent condition by which =2. Moreover, we show a sufficient condition for a multigraph to have Nn2>Nn-1Nn+1. As special cases of the sufficient condition, we show that if G contains at least +1 multiple edges between some pair of vertices, or if its underlying simple graph has no cycle with length more than 4, then Nn2>Nn-1Nn+1.

Effect of the post-growth annealing on the morphology of a Ge mesa selectively grown on Si was studied from the viewpoint of near-infrared photodiode applications. By ultrahigh-vacuum chemical vapor deposition, Ge mesas were selectively grown at 600 on Si (001) substrates partially covered with SiO2 masks. The as-grown Ge mesas showed trapezoidal cross-sections having a top (001) surface and {311} sidewall facets, as similar to previous reports. However, after the subsequent post-growth annealing at ~800 in the ultrahigh-vacuum chamber, the mesas were deformed into rounded shapes having a depression at the center and mounds near the edges. Such a deformation cannot be observed for the samples annealed once after cooled and exposed to the air. The residual hydrogen atoms on the Ge surface from the germane (GeH4) decomposition is regarded as a trigger to the observed morphological instability, while the final mesa shape is determined in order to minimize a sum of the surface and/or strain energies.

When a store sells items to customers, the store wishes to decide the prices of the items to maximize its profit. If the store sells the items with low (resp. high) prices, the customers buy more (resp. less) items, which provides less profit to the store. It would be hard for the store to decide the prices of items. Assume that a store has a set V of n items and there is a set C of m customers who wish to buy those items. The goal of the store is to decide the price of each item to maximize its profit. We refer to this maximization problem as an item pricing problem. We classify the item pricing problems according to how many items the store can sell or how the customers valuate the items. If the store can sell every item i with unlimited (resp. limited) amount, we refer to this as unlimited supply (resp. limited supply). We say that the item pricing problem is single-minded if each customer j ∈ C wishes to buy a set ej ⊆ V of items and assigns valuation w(ej) ≥ 0. For the single-minded item pricing problems (in unlimited supply), Balcan and Blum regarded them as weighted k-hypergraphs and gave several approximation algorithms. In this paper, we focus on the (pseudo) degree of k-hypergraphs and the valuation ratio, i.e., the ratio between the smallest and the largest valuations. Then for the single-minded item pricing problems (in unlimited supply), we show improved approximation algorithms (for k-hypergraphs, general graphs, bipartite graphs, etc.) with respect to the maximum (pseudo) degree and the valuation ratio.

A new subfamily of sequences with optimal correlation properties is constructed for the generalized Kasami set. A lower bound on the linear span is established. It is proved that with suitable choices of parameters, this subfamily has exponentially larger linear spans than either No sequences or TN sequences. A class of sequences with ideal autocorrelation is also proved to have large linear span.

In this paper, we propose a method that recovers a smooth high-resolution image from several blurred and roughly quantized low-resolution images. For compensation of the quantization effect we introduce measurements of smoothness, Huber function that is originally used for suppression of block noises in a JPEG compressed image [Schultz & Stevenson '94] and a smoothed version of total variation. With a simple operator that approximates the convex projection onto constraint set defined for each quantized image [Hasegawa et al. '05], we propose a method that minimizes these cost functions, which are smooth convex functions, over the intersection of all constraint sets, i.e. the set of all images satisfying all quantization constraints simultaneously, by using hybrid steepest descent method [Yamada & Ogura '04]. Finally in the numerical example we compare images derived by the proposed method, Projections Onto Convex Sets (POCS) based conventinal method, and generalized proposed method minimizing energy of output of Laplacian.