This paper presents a knowledge acquisition method using sentence topics for question answering. We define templates for information extraction by the Korean concept network semi-automatically. Moreover, we propose the two-phase information extraction model by the hybrid machine learning such as maximum entropy and conditional random fields. In our experiments, we examined the role of sentence topics in the template-filling task for information extraction. Our experimental result shows the improvement of 18% in F-score and 434% in training speed over the plain CRF-based method for the extraction task. In addition, our result shows the improvement of 8% in F-score for the subsequent QA task.

Previous approaches for modeling Intrusion Detection System (IDS) have been on twofold: improving detection model(s) in terms of (i) feature selection of audit data through wrapper and filter methods and (ii) parameters optimization of detection model design, based on classification, clustering algorithms, etc. In this paper, we present three approaches to model IDS in the context of feature selection and parameters optimization: First, we present Fusion of Genetic Algorithm (GA) and Support Vector Machines (SVM) (FuGAS), which employs combinations of GA and SVM through genetic operation and it is capable of building an optimal detection model with only selected important features and optimal parameters value. Second, we present Correlation-based Hybrid Feature Selection (CoHyFS), which utilizes a filter method in conjunction of GA for feature selection in order to reduce long training time. Third, we present Simultaneous Intrinsic Model Identification (SIMI), which adopts Random Forest (RF) and shows better intrusion detection rates and feature selection results, along with no additional computational overheads. We show the experimental results and analysis of three approaches on KDD 1999 intrusion detection datasets.

This paper proposes a task scheduling approach for reliable cache architectures (RCAs) of multiprocessor systems. The RCAs dynamically switch their operation modes for reducing the usage of vulnerable SRAMs under real-time constraints. A mixed integer programming model has been built for minimizing vulnerability under real-time constraints. Experimental results have shown that our task scheduling approach achieved 47.7-99.9% less vulnerability than a conventional one.

The single-hidden-layer feedforward neural networks (SLFNs) are frequently used in machine learning due to their ability which can form boundaries with arbitrary shapes if the activation function of hidden units is chosen properly. Most learning algorithms for the neural networks based on gradient descent are still slow because of the many learning steps. Recently, a learning algorithm called extreme learning machine (ELM) has been proposed for training SLFNs to overcome this problem. It randomly chooses the input weights and hidden-layer biases, and analytically determines the output weights by the matrix inverse operation. This algorithm can achieve good generalization performance with high learning speed in many applications. However, this algorithm often requires a large number of hidden units and takes long time for classification of new observations. In this paper, a new approach for training SLFNs called least-squares extreme learning machine (LS-ELM) is proposed. Unlike the gradient descent-based algorithms and the ELM, our approach analytically determines the input weights, hidden-layer biases and output weights based on linear models. For training with a large number of input patterns, an online training scheme with sub-blocks of the training set is also introduced. Experimental results for real applications show that our proposed algorithm offers high classification accuracy with a smaller number of hidden units and extremely high speed in both learning and testing.

P2P DHT (Peer-to-Peer Distributed Hash Table) is one of typical techniques for realizing an efficient management of shared resources distributed over a network and a keyword search over such networks in a fully distributed manner. In this paper, we propose a new method for supporting conjunctive queries in P2P DHT. The basic idea of the proposed technique is to share a global information on past trials by conducting a local caching of search results for conjunctive queries and by registering the fact to the global DHT. Such a result caching is expected to significantly reduce the amount of transmitted data compared with conventional schemes. The effect of the proposed method is experimentally evaluated by simulation. The result of experiments indicates that by using the proposed method, the amount of returned data is reduced by 60% compared with conventional P2P DHT which does not support conjunctive queries.

This paper deals with a TM plane wave reflection and transmission from a one-dimensional random slab with stratified fluctuation by means of the stochastic functional approach. Based on a previous manner [IEICE Trans. Electron. E88-C, 4, pp.713-720, 2005], an explicit form of the random wavefield is obtained in terms of a Wiener-Hermite expansion with approximate expansion coefficients (Wiener kernels) under small fluctuation. The optical theorem and coherent reflection coefficient are illustrated in figures for several physical parameters. It is then found that the optical theorem by use of the first two or three order Wiener kernels holds with good accuracy and a shift of Brewster's angle appears in the coherent reflection.

Energy consumption is a major concern in embedded computing systems. Several studies have shown that cache memories account for 40% or more of the total energy consumed in these systems. Active power used to be the primary contributor to total power dissipation of CMOS designs, but with the technology scaling, the share of leakage in total power consumption of digital systems continues to grow. Moreover, temperature is another factor that exponentially increases the leakage current. In this paper, we show the effect of temperature on the optimal (minimum-energy-consuming) cache configuration for low energy embedded systems. Our results show that for a given application and technology, the optimal cache size moves toward smaller caches at higher temperatures, due to the larger leakage. Consequently, a Temperature-Aware Configurable Cache (TACC) is an effective way to save energy in finer technologies when the embedded system is used in different temperatures. Our results show that using a TACC, up to 61% energy can be saved for instruction cache and 77% for data cache compared to a configurable cache that has been configured for only the corner-case temperature (100). Furthermore, the TACC also enhances the performance by up to 28% for the instruction cache and up to 17% for the data cache.

Learning for boltzmann machines deals with each state individually. If given data is categorized, the probabilities have to be distributed to each state, not to each catetory. We propose boltzmann machines identifying the states in the same categories. Boltzmann machines with hidden units are the special cases. Boltzmann learning and em algorithm are effective learning methods for boltzmann machines. We solve boltzmann learning and em algorithm for the proposed models.

Error concealment at a decoder is an efficient method to reduce the degradation of visual quality caused by channel errors. In this paper, we propose a novel spatio-temporal error concealment algorithm based on the spatial-temporal fading (STF) scheme which has been recently introduced. Although STF achieves good performance for the error concealment, several drawbacks including blurring still remain in the concealed blocks. To alleviate these drawbacks, in the proposed method, hybrid approaches with adaptive weights are proposed. First, the boundary matching algorithm and the decoder motion vector estimation which are well-known temporal error concealment methods are adaptively combined to compensate for the defect of each other. Then, an edge preserved method is utilized to reduce the blurring effects caused by the bilinear interpolation for spatial error concealment. Finally, two concealed results obtained by the hybrid spatial and temporal error concealment are pixel-wisely blended with adaptive weights. Experimental results exhibit that the proposed method outperforms conventional methods including STF in terms of the PSNR performance as well as subjective visual quality, and the computational complexity of the proposed method is similar to that of STF.

We have attempted to recognize reverberant speech using a novel speech recognition system that depends on not only the spectral envelope and amplitude modulation but also frequency modulation. Most of the features used by modern speech recognition systems, such as MFCC, PLP, and TRAPS, are derived from the energy envelopes of narrowband signals by discarding the information in the carrier signals. However, some experiments show that apart from the spectral/time envelope and its modulation, the information on the zero-crossing points of the carrier signals also plays a significant role in human speech recognition. In realistic environments, a feature that depends on the limited properties of the signal may easily be corrupted. In order to utilize an automatic speech recognizer in an unknown environment, using the information obtained from other signal properties and combining them is important to minimize the effects of the environment. In this paper, we propose a method to analyze carrier signals that are discarded in most of the speech recognition systems. Our system consists of two nonlinear discriminant analyzers that use multilayer perceptrons. One of the nonlinear discriminant analyzers is HATS, which can capture the amplitude modulation of narrowband signals efficiently. The other nonlinear discriminant analyzer is a pseudo-instantaneous frequency analyzer proposed in this paper. This analyzer can capture the frequency modulation of narrowband signals efficiently. The combination of these two analyzers is performed by the method based on the entropy of the feature introduced by Okawa et al. In this paper, in Sect. 2, we first introduce pseudo-instantaneous frequencies to capture a property of the carrier signal. The previous AM analysis method are described in Sect. 3. The proposed system is described in Sect. 4. The experimental setup is presented in Sect. 5, and the results are discussed in Sect. 6. We evaluate the performance of the proposed method by continuous digit recognition of reverberant speech. The proposed system exhibits considerable improvement with regard to the MFCC feature extraction system.

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.

Robust automatic language identification (LID) is the task of identifying the language from a short utterance spoken by an unknown speaker. The mainstream approaches include parallel phone recognition language modeling (PPRLM), support vector machine (SVM) and the general Gaussian mixture models (GMMs). These systems map the cepstral features of spoken utterances into high level scores by classifiers. In this paper, in order to increase the dimension of the score vector and alleviate the inter-speaker variability within the same language, multiple data groups based on supervised speaker clustering are employed to generate the discriminative language characterization score vectors (DLCSV). The back-end SVM classifiers are used to model the probability distribution of each target language in the DLCSV space. Finally, the output scores of back-end classifiers are calibrated by a pair-wise posterior probability estimation (PPPE) algorithm. The proposed language identification frameworks are evaluated on 2003 NIST Language Recognition Evaluation (LRE) databases and the experiments show that the system described in this paper produces comparable results to the existing systems. Especially, the SVM framework achieves an equal error rate (EER) of 4.0% in the 30-second task and outperforms the state-of-art systems by more than 30% relative error reduction. Besides, the performances of proposed PPRLM and GMMs algorithms achieve an EER of 5.1% and 5.0% respectively.

Modern applications of pattern recognition generate very large amounts of data, which require large computational effort to process. However, the majority of the methods intended for large-scale problems aim to merely adapt standard classification methods without considering if those algorithms are appropriated for large-scale problems. CombNET-II was one of the first methods specifically proposed for such kind of a task. Recently, an extension of this model, named CombNET-III, was proposed. The main modifications over the previous model was the substitution of the expert networks by Support Vectors Machines (SVM) and the development of a general probabilistic framework. Although the previous model's performance and flexibility were improved, the low accuracy of the gating network was still compromising CombNET-III's classification results. In addition, due to the use of SVM based experts, the computational complexity is higher than CombNET-II. This paper proposes a new two-layered gating network structure that reduces the compromise between number of clusters and accuracy, increasing the model's performance with only a small complexity increase. This high-accuracy gating network also enables the removal the low confidence expert networks from the decoding procedure. This, in addition to a new faster strategy for calculating multiclass SVM outputs significantly reduced the computational complexity. Experimental results of problems with large number of categories show that the proposed model outperforms the original CombNET-III, while presenting a computational complexity more than one order of magnitude smaller. Moreover, when applied to a database with a large number of samples, it outperformed all compared methods, confirming the proposed model's flexibility.

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.

In this paper, a novel cache conscious indexing technique based on space partitioning trees is proposed. Many researchers investigated efficient cache conscious indexing techniques which improve retrieval performance of in-memory database management system recently. However, most studies considered data partitioning and targeted fast information retrieval. Existing data partitioning-based index structures significantly degrade performance due to the redundant accesses of overlapped spaces. Specially, R-tree-based index structures suffer from the propagation of MBR (Minimum Bounding Rectangle) information by updating data frequently. In this paper, we propose an in-memory space partitioning index structure for optimal cache utilization. The proposed index structure is compared with the existing index structures in terms of update performance, insertion performance and cache-utilization rate in a variety of environments. The results demonstrate that the proposed index structure offers better performance than existing index structures.

Most learning-based super-resolution methods neglect the illumination problem. In this paper we propose a novel method to combine blind single-frame super-resolution and shadow removal into a single operation. Firstly, from the pattern recognition viewpoint, blur identification is considered as a classification problem. We describe three methods which are respectively based on Vector Quantization (VQ), Hidden Markov Model (HMM) and Support Vector Machines (SVM) to identify the blur parameter of the acquisition system from the compressed/uncompressed low-resolution image. Secondly, after blur identification, a super-resolution image is reconstructed by a learning-based method. In this method, Logarithmic-wavelet transform is defined for illumination-free feature extraction. Then an initial estimation is obtained based on the assumption that small patches in low-resolution space and patches in high-resolution space share a similar local manifold structure. The unknown high-resolution image is reconstructed by projecting the intermediate result into general reconstruction constraints. The proposed method simultaneously achieves blind single-frame super-resolution and image enhancement especially shadow removal. Experimental results demonstrate the effectiveness and robustness of our method.

This paper proposes a method (Group-Linking Method) that has control over the complexity of the sequential function to construct Finite Memory Machines with minimal order--the machines have the largest number of states based on their memory taps. Finding a machine with maximum number of states is a nontrivial problem because the total number of machines with memory order k is (256)2k-2, a pretty large number. Based on the analysis of Group-Linking Method, it is shown that the amount of data necessary to reconstruct an FMM is the set of strings not longer than the depth of the machine plus one, which is significantly less than that required for traditional greedy-based machine learning algorithm. Group-Linking Method provides a useful systematic way of generating unified benchmarks to evaluate the capability of machine learning techniques. One example is to test the learning capability of recurrent neural networks. The problem of encoding finite state machines with recurrent neural networks has been extensively explored. However, the great representation power of those networks does not guarantee the solution in terms of learning exists. Previous learning benchmarks are shown to be not rich enough structurally in term of solutions in weight space. This set of benchmarks with great expressive power can serve as a convenient framework in which to study the learning and computation capabilities of various network models. A fundamental understanding of the capabilities of these networks will allow users to be able to select the most appropriate model for a given application.

Recently, many techniques have been proposed to improve speaker identification in noise environments. Among these techniques, we consider the feature recombination technique for the multi-band approach in noise robust speaker identification. The conventional feature recombination technique is very effective in the band-limited noise condition, but in broad-band noise condition, the conventional feature recombination technique does not provide notable performance improvement compared with the full-band system. Even though the speech is corrupted by the broad-band noise, the degree of the noise corruption on each sub-band is different from each other. In the conventional feature recombination for speaker identification, all sub-band features are used to compute multi-band likelihood score, but this likelihood computation does not use a merit of multi-band approach effectively, even though the sub-band features are extracted independently. Here we propose a new technique of sub-band likelihood computation with sub-band weighting in the feature recombination method. The signal to noise ratio (SNR) is used to compute the sub-band weights. The proposed sub-band-weighted likelihood computation makes a speaker identification system more robust to noise. Experimental results show that the average error reduction rate (ERR) in various noise environments is more than 24% compared with the conventional feature recombination-based speaker identification system.

An ultra-deep polymer cavity structure exposed using deep X-ray lithography is used as a template for metal electroforming to produce a 24-GHz cavity resonator. The metal cavity is 1.8 mm deep and has impressive structure, including extremely vertical and smooth sidewalls, resulting in low conductor loss. The measured resonator has an unloaded quality factor of above 1800 at a resonant frequency of 23.89 GHz.

In this paper, we describe a method of applying Collaborative Filtering with a Machine Learning technique to predict users' preferences for clothes on online shopping malls when user history is insufficient. In particular, we experiment with methods of predicting missing values, such as mean value, SVD, and support vector regression, to find the best method and to develop and utilize a unique feature vector model.