The important thing in order that the CRT still shines brightly for the coming digital era might be to realize a super-lightweight and ultra thin CRT, keeping its advantages such as high image quality and low cost. One of the reasons that CRT is heavy and fragile is due to the high density of CRT glass resulted from its high X-ray absorption ability. However, it might become an impetus for realizing such an innovative CRT by introducing new CRT glass with lower density and lower brittleness. This could be possible if a new electron gun with excellent focusing characteristics even at very low accelerating voltages would be developed. In addition, we have had some advances in strengthening glass material such as field-assisted chemical tempering and improvement of the mechanical strength of a crystalline solder glass. These also will produce a possibility to innovate CRTs.

抄録全体を表示

Stripe ribs make address discharge more stable than box ribs due presumably to the priming particles provided from vertically neighboring cells. Discharge deactivation film (DDF) of low γ_i material, which selectively covers an MgO surface so as to fold the relative bus lines, improves the address discharge response dramatically because DDF pushes the address discharge area closer to the surface discharge gap. A TiO₂ under-layer for phosphor also significantly improves the response. Presumably the electrification characteristic of the layer, which may depends on the surface treatment onto the TiO₂ grains, contributes much to the formation of an electric field for generating face-to-face address discharge. Using these in our 46-in. HD-PDP, Xe15% (by volume) is acceptable with respect to practical driving for TV use if the dielectric layer is arranged thinly to prevent extremely high sustain voltage. Thus white peak luminance and luminous efficiency of 1220 cd/m² and 2.16 lm/W, respectively, can be achieved.

抄録全体を表示

Autonomous concurrent objects can be regarded as message receivers rather than labeled records by message senders, each having object references. The subtype relation on reference types can be determined by judging how various messages can be understood. Based on this idea, this paper presents a type system and compilation techniques for concurrent objects, implementing efficient pattern matching of first class messages, where the look-up can be performed as direct indexing of a small table, and the table size can be bounded by a number slightly more than the number of injection tags which the message value may have.

抄録全体を表示

Inductive inference gives us a theoretical model of concept learning from examples. In this paper, we study refutably and reliably inductive inference of recursive real-valued functions. First we introduce the new criteria RealRefEx for refutable inference and RealRelEx for reliable inference. Then, we compare these two criteria with RealEx for identification in the limit, RealFin for learning finitely and RealNum¡ for learning by enumeration that have been already introduced in the previous works, and investigate their interaction. In particular, we show that RealRefEx and RealRelEx are closed under union, as similar as the criteria RefEx and RelEx for inductive inference of recursive functions.

抄録全体を表示

This paper proposes a middle-grain approach to construct hybrid MPI-OpenMP solutions for SMP clusters from an existing MPI algorithm. Experiments on different cluster platforms show that our solutions exceed the solutions that are based on the de-facto MPI model in most cases, and occasionally by as much as 40% of performance. We also prove an automatic outperformance of a thread-to-thread communication model over a traditional process-to-process communication model in hybrid solutions. In addition, the paper performs a detailed analysis on the hardware and software factors affecting the performance of MPI in comparison to hybrid models.

抄録全体を表示

Inductive Logic Programming (ILP) becomes interesting when the expressive power of first-order representation provides comprehensibility to learning result and capability to handle more complex data consisting of their relations. Nevertheless, the bottleneck for learning first-order theory is enormous hypothesis search space which causes inefficient performance by the existing learning approaches compared to the propositional approaches. This paper introduces an improved ILP approach capable of handling more efficiently a kind of data called multiple-part data, i.e., one instance of data consists of several parts as well as relations among parts. This approach tries to find hypothesis describing class of each training example by using both individual and relational characteristics of its part which is similar to finding common substructures among the complex relational instances. The multiple-part data can be found in various domains especially on Structure-Activity Relationship (SAR) studies which aim to generate hypotheses describing activities or characteristics of chemical compounds from their own structures. Each compound is composed of atoms as parts, and various kinds of bond as relations among atoms. We then apply the proposed algorithm for SAR studies by conducting experiments on two real-world datasets: mutagenicity in nitroaromatic compounds and dopamine antagonist compounds. The experiment results were compared to the previous approaches in order to show the performance of proposed approach.

抄録全体を表示

We propose a behavioral model of web applications, called `Web Automata', based on the MVC(Model View and Control) model architecture. The MVC model architecture separates design concerns to improve the overall software quality. Since the architecture only defines the abstract outline of the configurations, there is a broad gap between web application codes and their behavioral properties.
We model the behavior of a web application with dynamic contents as an extension of links-automata proposed by Stotts et al. with the constraint-logic feature of the Extended Finite Automata, EFA for short, by Sarna-Starosa and Ramakrishna. As extended in the model checking techniques, we view a web application as a data-independent system, where variables appearing in link parameters and form inputs are attached to each page.
We present a testing framework for web applications based on the behavioral model. We show it provides reasonable testing criteria for web applications when we focus on the loops in the Web automaton model. We apply our framework to the Jakarta Struts by presenting the extended configuration schema of Struts in order to describe the web automata directly.

抄録全体を表示

The pursuit of instruction-level parallelism using more transistors produces diminishing returns and also increases power dissipation of general purpose processors. This paper studies a chip multi-processor (CMP) with smaller processor cores as a means to achieve high aggregate throughput and improved energy efficiency. The benefit of this design approach increases as the number of cores on a chip increases, as enabled by semiconductor process scaling. The feasibility of a processor core 40% of the size of a baseline high performance processor that delivers about 70% of its performance is shown. The CMP populated by smaller cores to fill the same silicon area delivers 2.3 times higher performance in transaction processing represented by TPC-C benchmarks than the baseline processor scaled into the same technology. The CMP also achieves 38% higher energy efficiency.

抄録全体を表示

Detecting outliers is an important problem, in applications such as fraud detection, financial analysis, health monitoring and so on. It is typical of most such applications to possess high dimensional datasets. Many recent approaches detect outliers according to some reasonable, pre-defined concepts of an outlier (e.g., distance-based, density-based, etc.). Most of these concepts are proximity-based which define an outlier by its relationship to the rest of the data. However, in high dimensional space, the data becomes sparse which implies that every object can be regarded as an outlier from the point of view of similarity. Furthermore, a fundamental issue is that the notion of which objects are outliers typically varies between users, problem domains or, even, datasets. In this paper, we present a novel solution to this problem, by detecting outliers based on user examples for high dimensional datasets. By studying the behavior of projections of such a few outlier examples in the dataset, the proposed method discovers the hidden view of outliers and picks out further objects that are outstanding in the projection where the examples stand out greatly. Our experiments on both real and synthetic datasets demonstrate the ability of the proposed method to detect outliers that match users' intentions.

抄録全体を表示

In the Shamir (t, n)-threshold scheme, the dealer constructs a random polynomial f(x) ∈ GF(p)[x] of degree at most t-1 in which the constant term is the secret K ∈ GF(p). However, if the chosen polynomial f(x) is of degree less than t-1, then a conspiracy of any t-1 participants can reconstruct the secret K;on the other hand, if the degree of f(x) is greater than t-1, then even t participants can not reconstruct the secret K properly. To prevent these from happening, the degree of the polynomial f(x) should be exactly equal to t-1 if the dealer claimed that the threshold of this scheme is t. There also should be some ways for participants to verify whether the threshold is exactly t or not. A few known verifiable threshold schemes provide such ability but the securities of these schemes are based on some cryptographic assumptions. The purpose of this paper is to propose some threshold-verification protocols for the Shamir (t, n)-threshold scheme from the viewpoint of unconditional security.

抄録全体を表示

When a hard drive (HDD) is recycled, it is recommended that all files on the HDD are repeatedly overwritten with random strings for protecting their confidentiality. However, it takes a long time to overwrite them. This problem is solved by applying the all-or-nothing transform (AONT) to the filesystem of the HDD. To use the HDD economically, it is desirable to use a length-preserving AONT (LP-AONT). Whereas previous AONTs cause the increase of size of a file, and no LP-AONT is secure under previous security definitions. However, it does not mean that the LP-AONT is useless;previous security definitions are too strict in practical applications. Then, by introducing the ambiguity of a message, we propose more practical security definitions of the AONT. We also show the secure implementation of the LP-AONT under the proposed security definitions. The analysis shows that our implementation is nearly optimal in terms of the success probability of an adversary. It means that the ambiguity of one message block allows us to construct the LP-AONT as secure as previous AONTs.

抄録全体を表示

This paper presents a method to evaluate the risk of information leakage in software processes for security-sensitive applications. A software process is modeled as a series of sub-processes, each of which produces new work products from input products. Since a process is conducted usually by multiple developers, knowledge of work products is shared among the developers. Through the collaboration, a developer may share with others the knowledge of products that are not related to the process. We capture the transfer of such irrelevant product knowledge as information leakage in a software process. In this paper, we first formulate the problem of information leakage by introducing a formal software process model. Then, we propose a method to derive the probability that each developer d knows each work product p at a given process of software development. The probability reflects the possibility that someone leaked the knowledge of p to d. We also conduct three case studies to show the applicability of leakage to practical settings. In the case studies, we evaluate how the risk of information leakage is influenced by the collaboration among developers, the optimal developer assignment and the structure of the software process. As a result, we show that the proposed method provides a simple yet powerful means to perform quantitative analysis on information leakage in a security-sensitive software process.

抄録全体を表示

In Java programs, it is difficult to protect intellectual property rights and secret information in untrusted environments, since they are easy to decompile and reverse engineer. Consequently realization of software obfuscation becomes increasingly important. Unfortunately previous software obfuscation techniques share a major drawback that they do not have a theoretical basis and thus it is unclear how effective they are. Therefore we shall propose new software obfuscation techniques for Java in this paper. Our obfuscation techniques take advantage of features of object-oriented languages, and they drastically reduce the precision of points-to analysis of the programs. We show that determining precise points-to analysis in obfuscated programs is NP-hard and the fact provides a theoretical basis for our obfuscation techniques. Furthermore, in this paper we present some empirical experiments, whereby we demonstrate the effectiveness of our approaches.

抄録全体を表示

The notion of advised computation was introduced by Karp and Lipton to represent non-uniform complexity in terms of Turing machines. Since then, advised computation has been one of the basic concepts of computational complexity. Recently, the research of advised computation has been originated also in the field of quantum computing. This paper reviews the study of advised quantum computation.

抄録全体を表示

We study a distinction problem of two heat baths with given information from ancilla (quantum or classical two level system) which is interacting with the heat baths. Analysing dynamics of ancilla system in a Markov approximation region, we show that the process in which“quantum ancilla” acquires the quantum distinguishability is described as a bifurcation process from the time evolution of classical distinguishability.

抄録全体を表示

Quantum automata have been studied as simple quantum computation models. They can be considered models of small (or restricted)quantum computers. In this paper, we give descriptions of several kinds of quantum automata and show their power in comparison to their classical counterparts. We also give descriptions of quantum automata that have additional classical computational resources. Introducing classical computational resources can enhance the power of quantum automata, since this approach relaxes such restrictions as reversible state transitions.

抄録全体を表示

This paper reviews researches on quantum oracle computations when oracles are not perfect, i.e., they may return wrong answers. We call such oracles biased oracles, and discuss the formal model of them. Then we provide an intuitive explanation how quantum search with biased oracles by Høyer, et al.(2003) works. We also review the method, by Buhrman, et al.(2005), to obtain all the answers of a quantum biased oracle without any overhead compared to the perfect oracle case. Moreover, we discuss two special cases of quantum biased oracles and their interesting properties, which are not found in the classical corresponding cases. Our discussion implies that the model of quantum biased oracle adopted by the existing researches is natural.

抄録全体を表示

After Shor's discovery of an efficient quantum algorithm for integer factoring, hidden subgroup problems play a central role in developing efficient quantum algorithms. In spite of many intensive studies, no efficient quantum algorithms are known for hidden subgroup problems for many non-Abelian groups. Of particular interest are the hidden subgroup problems for the symmetric group and for the dihedral group, because an efficient algorithm for the former implies an efficient solution to the graph isomorphism problem, and that for the latter essentially solves a certain lattice-related problem whose hardness is assumed in cryptography. This paper focuses on the latter case and gives a comprehensive survey of known facts related to the dihedral hidden subgroup problem.

抄録全体を表示

In this paper we propose a new framework for transformations of XML documents based on an extension of regular expression type, called incomplete regular expression type. Incomplete regular expression type is a restricted second-order extension of regular expression (RE) type: An incomplete RE type can be applied to arbitrary RE types as its arguments. It is accomplished by introducing holes to types. We give a version of second-order rewrite systems, designed founded on our type system. Matching between a subterm with the left-hand side of rewrite rules yields a substitution that bind second-order terms to the variables of incomplete type. This feature is practically useful when we want to reuse “the upper parts” of XML documents in the transformation. We demonstrate that the introduction of incomplete regular expression type allows programmers much flexibility. In this paper, we mainly concentrate on the type theoretic issues and give some properties concerning for incomplete types. We also show the type preservation in rewriting.

抄録全体を表示

Ambient calculus is a process algebra developed for describing mobile processes. Ambients represent the substances of movement and the fields of the ambients themselves. Having this hierarchy, it can model various kinds of mobile computation. Equational relation for ambient calculus “Contextual Equivalence” were proposed regarding the names of ambients observed from the environment. This relation is, however, not strong as “testing equivalence” so that it can identify the processes which have different properties. This paper proposes equational relations for ambient calculus by which we can distinguish processes that the existing equivalence identifies.

抄録全体を表示

New modulo m multipliers with a radix-two signed-digit (SD) number arithmetic is presented by using a modified Booth recoding method. To implement a modulo m multiplication, we usually generate modulo m partial products, then perform modulo m sum of them. In this paper, a new Booth recoding method is proposed to convert a radix-two SD number into a recoded SD (RSD) number in parallel. In the RSD number representation, there are no (1, 1)and (-1, -1) at any two-digit position. Thus, by using the RSD converted, the modulo m partial products can be cut from n into n/2 for an n × n modulo m multiplication. Parallel and serial modulo m multipliers have been designed by using the SD number arithmetic and the proposed Booth recoding. Compared to the former work, the area for VLSI implementation of the parallel modulo m multiplier is reduced to 80% from the original design, and the speed performance of the serial multiplier is improved up to twice by using the Booth recoding. The implementation method of the proposed Booth modulo m multipliers has been verified by a gate level simulation.

抄録全体を表示

A three-dimensional, relativistic, electromagnetic particle simulation code is developed using the “Exact Charge Conservation Scheme”on a vector-parallel supercomputer. This scheme is a method for calculating current densities. Applying this method, correction of electric fields is not needed. In this paper, some techniques to optimize the above scheme for a vector-parallel supercomputer are shown. The method for vectorization and parallelization in shared memory and in distributed memories are discussed. The code is written in Fortran90 and High Performance Fortran (HPF). Examination of this code is also made.

抄録全体を表示

We propose some PAC like settings for a learning problem of a sub-class of linear languages, and show its polynomial time learnability in each of our settings. Here, the sub-class of linear languages is newly defined, and it includes the class of regular languages and the class of even linear languages. We show a polynomial time learning algorithm in either of the following settings with a fixed but unknown probability distribution for examples.(1) The first case is when the learner can use randomly drawn examples, membership queries, and a set of representative samples.(2) The second case is when the learner can use randomly drawn examples, membership queries, and both of the size of a grammar which can generate the target language and d. Where d is the probability such that the rarest rule in the target grammar occurs in the derivation of a randomly drawn example. In each case, for the target language L_t, the hypothesis L_hsatisfies thatPr[P(L_h Δ L_t) ≤ ε] ≥ 1 - δ for the error parameter 0 < ε ≤ 1 and the confidential parameter 0 < δ ≤ 1.

抄録全体を表示

The notion of the tree edit distance provides a unifying framework for measuring distance and finding approximate common patterns between two trees. A diversity of tree edit distance measures have been proposed to deal with tree related problems, such as minor containment, maximum common subtree isomorphism, maximum common embedded subtree, and alignment of trees. These classes of problems are characterized by the conditions of the tree mappings, which specify how to associate the nodes in one tree with the nodes in the other. In this paper, we study the declarative semantics of edit distance measures based on the tree mapping. In prior work, the edit distance measures have been not well-formalized. So the relationship among various algorithms based on the tree edit distance has hardly been studied. Our framework enables us to study the relationship. By using our framework, we reveal the declarative semantics of the alignment of trees, which has remained unknown in prior work.

抄録全体を表示

It is common that a word in any natural language has often more than one meaning/sense. A word sense disambiguation (WSD) system is designed to determine which one of the senses of a polysemous word is invoked in a particular context around the word. We propose methods to disambiguate senses of polysemous words by using Naive Bayesian classifier method. A few sets of experiment data were taken from Kompas daily newspaper homepage and used for the system construction. We modified the original algorithm of Naive Bayesian method to apply it to the Indonesian language analysis. The experiments showed that our system achieved good accuracies (73-99%).

抄録全体を表示

An algorithm for reconstructing photorealistic 3D model from multiple-view images is proposed. The idea is based on the surface light field approach. In the algorithm, a geometric model is reconstructed as a visual hull using an image-based multi-pass algorithm we have developed, then the hull is represented as a quadrilateral-meshed surface. Next, colors of input images are assigned onto each vertex according to viewing directions using a new data structure we have developed. The structure is a hexagonal tessellation based on expansion and replacement of a buckyball. Finally, the hexagonal tessellation is represented as a hexagonal image whose pixels represent colors of corresponding input images. Experimental results for real objects show that 3D data can be successfully generated automatically in a short time and that photorealistic data can be viewed from arbitrary viewpoints even for objects with reflective or translucent surfaces.

抄録全体を表示

Research on human-robot interaction is getting an increasing amount of attention. Since most research has dealt with communication between one robot and one person, quite few researchers have studied communication between a robot and multiple people. This paper presents a method that enables robots to communicate with multiple people using the “selection priority of the interactive partner” based on the concept of Proxemics. In this method, a robot changes active sensory-motor modalities based on the interaction distance between itself and a person. Our method was implemented into a humanoid robot, SIG2. SIG2 has various sensory-motor modalities to interact with humans. A demonstration of SIG2 showed that our method selected an appropriate interaction partner during interaction with multiple people.

抄録全体を表示

If a dialog system can respond to the user as reasonably as a human, the interaction will become smoother. Timing of the response such as back-channels and turn-taking plays an important role in such a smooth dialog as in human-human interaction. We developed a response timing generator for such a dialog system. This generator uses a decision tree to detect the timing based on the features coming from some prosodic and linguistic information. The timing generator decides the action of the system at every 100 ms during the user's pause. In this paper, we describe a robust spoken dialog system using the timing generator. Subjective evaluation proved that almost all of the subjects experienced a friendly feeling from the system.

抄録全体を表示

A novel approach to human-robot collaboration based on quasi-symbolic expressions is proposed. The target task is navigation in which a person with his or her eyes covered and a humanoid robot collaborate in a context-dependent manner. The robot uses a recurrent neural net with parametric bias (RNNPB) model to acquire the behavioral primitives, which are sensory-motor units, composing the whole task. The robot expresses the PB dynamics as primitives using symbolic sounds, and the person influences these dynamics through tactile sensors attached to the robot. Experiments with six participants demonstrated that the level of influence the person has on the PB dynamics is strongly related to task performance, the person's subjective impressions, and the prediction error of the RNNPB model (task stability). Simulation experiments demonstrated that the subjective impressions of the correspondence between the utterance sounds (the PB values) and the motions were well reproduced by the rehearsal of the RNNPB model.

抄録全体を表示

In this paper, we propose a novel kernel computation algorithm between time-series human motion data for online action recognition. The proposed kernel is based on probabilistic models called switching linear dynamics (SLDs). SLD is one of the powerful tools for tracking, analyzing and classifying human complex time-series motion. The proposed kernel incorporates information about the latent variables in SLDs. The empirical evaluation using real motion data shows that a classifier using SVM with our proposed kernel has much better performance than the classifiers with some conventional kernel techniques. Another experimental result using kernel principal component analysis shows that the proposed kernel has excellent performance in extracting and separating different action categories, such as walking and running.

抄録全体を表示

This study is meant to contribute, from a psychological viewpoint, to the development of a “symbiotic” system, an intelligent system capable of “living with” human beings. We approach this by examining how people interact gesturally with each other, with a special focus on breathing movements. Within this general framework, the present paper reports two experiments conducted to examine the dynamics underlying intra- and inter-personal coordination of speech articulation, hand gesture movements, and breathing movements. The results reveal similarities as well as differences between intra- and inter-personal coordination, and we discuss their implications for existing theories of motor coordination, as well as for the development of human-machine symbiotic systems.

抄録全体を表示

We propose a new method of using multiple documents as evidence with decreased adding to improve the performance of question-answering systems. Sometimes, the answer to a question may be found in multiple documents. In such cases, using multiple documents to predict answers may generate better answers than using a single document. Our method therefore uses information from multiple documents, adding the scores of candidate answers extracted from various documents. However, because simply adding the scores can degrade the performance of question-answering systems, we add the scores with progressively decreasing weights to reduce the negative effect of simple adding. We carried out experiments using the Question-Answering Challenge (QAC) test collection. The results showed that our method produced a statistically significant improvement, with the degree of improvement ranging from 0.05 to 0.14. These results, and the fact that our method is simple and easy to use, indicate its potential feasibility and utility in question-answering systems. Experiments comparing our decreased adding method with several previously proposed methods that use multiple documents showed that our method was more effective than these other methods.

抄録全体を表示

We propose a neural preprocess approach for video-based gesture recognition system. Second-order neural network (SONN) and self-organizing map (SOM) are employed for extracting moving hand regions and for normalizing motion features respectively. The SONN is more robust to noise than frame difference technique. Obtained velocity feature vectors are translated into normalized feature space by the SOM with keeping their topology, and the transition of the activated node in the topological map is classified by DP matching. The topological nature of the SOM is quite suited to data normalization for the DP matching technique. Experimental results show that those neural networks effectively work on the gesture pattern recognition. The SONN shows its noise reduction ability for noisy backgrounds, and the SOM provides the robustness to spatial scaling of input images. The robustness of the SOM to spatial scaling is based on its robustness to velocity scaling.

抄録全体を表示

This paper describes a method to detect grammatical errors from a non-native speaker's utterance for a dialogue-based CALL (Computer Assisted Language Learning) system. For conversation exercises, several dialogue-based CALL systems were developed. However, one of the problems in conventional dialogue-based CALL systems is that a learner is usually assigned a passive role. The goal of our system is to allow a learner to compose his/her own sentences freely in a role-playing situation. One of the biggest problems in realizing the proposed system is that the learner's utterance inevitably contains pronunciation, lexical and grammatical errors. In this paper, we focus on the correction of the lexical and grammatical errors. To correct these errors, we propose two methods to detect lexical/grammatical errors in an utterance. The conventional methods are to write a grammar that accepts the errors manually. The proposed methods 1 and 2 use the `error rules' that are independent of the recognition grammar. The method 1 uses only correct system grammar and extends the recognition results using the `error rules'. The method 2 uses a general grammar (which does not consider the relationship between verb, particle and each noun) to recognize the learner's utterance and check acceptance of each N-best result and searches the learner's utterance. The grammar error detection experiment proved that the method 2 performs as well as the conventional method.

抄録全体を表示

A simpler distribution that fits empirical word distribution about as well as a negative binomial is the Katz K mixture. In the K mixture model, the basic assumption is that the conditional probabilities of repeats for a given word are determined by a constant decay factor that is independent of the number of occurrences which have taken place. However, the probabilities of the repeat occurrences are generally lower than the constant decay factor for the content-bearing words with few occurrences that have taken place. To solve this deficiency of the K mixture model, in-depth exploration of the characteristics of the conditional probabilities of repetitions, decay factors and their influences on modeling term distributions was conducted. Based on the results of this study, it appears that both ends of the distribution can be used to fit models. That is, not only can document frequencies be used when the instances of a word are few, but also tail probabilities (the accumulation of document frequencies). Both document frequencies for few instances of a word and tail probabilities for large instances are often relatively easy to estimate empirically. Therefore, we propose an effective approach for improving the K mixture model, where the decay factor is the combination of two possible decay factors interpolated by a function depending on the number of instances of a word in a document. Results show that the proposed model can generate a statistically significant better estimation of frequencies, especially the frequency estimation for a word with two instances in a document. In addition, it is shown that the advantages of this approach will become more evident in two cases, modeling the term distribution for the frequently used content-bearing word and modeling the term distribution for a corpus with a wide range of document length.

抄録全体を表示

We have proposed a K-means clustering based target tracking method, compared with the template matching, which can work robustly when tracking an object with hole through which the background can be seen (e.g., mosquito coil) (hereafter we call this problem as the background interfusionor the interfused background). This paper presents a new method for solving the drawbacks of the previous method, i.e., low speed, instability caused by the change of shape and size. Our new tracking model consists of a single target center, and a variable ellipse model for representing non-target pixels. The contributions of our new method are: 1) The original K-means clustering is replaced by a 2_∞-means clustering, and the non-target cluster center is adaptively picked up from the pixels on the ellipse. This modification reduces the number of distance computation and improves the stability of the target detection as well. 2) The ellipse parameters are adaptively adjusted according to the target detection result. This adaptation improves the robustness against the scale and shape changes of the target. Through the various experiments, we confirmed that our new method improves speed and robustness of our original method.

抄録全体を表示

This paper extends a greedy decoder for statistical machine translation (SMT), which searches for an optimal translation by using SMT models starting from a decoder seed, i.e., the source language input paired with an initial translation hypothesis. First, the outputs generated by multiple translation enginesare utilized as the initial translation hypotheses, whereby their variations reduce local optima problems inherent in the search. Second, a rescoring method based on the edit-distance between the initial translation hypothesis and the outputs of the decoder is used to compensate for problems of conventional greedy decoding solely based on statistical models. Our approach is evaluated for the translation of dialogues in the travel domain, and the results show that it drastically improves translation quality.

抄録全体を表示

Question classification is of crucial importance for question answering. In question classification, the accuracy of ML algorithms was found to significantly outperform other approaches. The two key issues in classification with a ML-based approach are classifier design and feature selection. Support Vector Machines is known to work well for sparse, high dimensional problems. However, the frequently used Bag-of-Words approach does not take full advantage of information contained in a question. To exploit this information we introduce three new feature types: Subordinate Word Category, Question Focus and Syntactic-Semantic Structure. As the results demonstrate, the inclusion of the new features provides higher accuracy of question classification compared to the standard Bag-of-Words approach and other ML based methods such as SVM with the Tree Kernel, SVM with Error Correcting Codes and SNoW. A classification accuracy of 85.6 % obtained using the three introduced feature types is, as of yet the highest reported in the literature, bringing error reduction of 27 % compared to the Bag-of-Words approach.

抄録全体を表示

To have an instructional plan guide the learning process is significant to various teaching styles and an important task in an ITS. Though various approaches have been used to tackle this task, the compelling need is for an ITS to improve on its own the plans established in a dynamic way. We hypothesize that the use of knowledge derived from student categories can significantly support the improvement of plans on the part of the ITS. This means that category knowledge can become effectors of effective plans. We have conceived a Category-based Self-improving Planning Module (CSPM) for an ITS tutor agent that utilizes the knowledge learned from learner categories to support self-improvement. The learning framework of CSPM employs unsupervised machine learning and knowledge acquisition heuristics for learning from experience. We have experimented on the feasibility of CSPM using recorded teaching scenarios.

抄録全体を表示

The Semantic Web and ontologies can be seen as a way to enable greater access not only to contents but also services on the Web. Users and software agents should be able to discover, invoke and compose Web services with a high degree of automation. However, current existing Web services are created by many different parties and are described based on different ontologies, which makes it difficult for retrieval systems to locate needed services. To facilitate the discovery of Web services in multi-ontology environments, we propose an approach to determine the semantic similarity of properties between different ontologies.

抄録全体を表示

A spurious timeout (STO) leads to an unnecessary go-back-N retransmission and throughput degradation, which negatively impacts the user's TCP performance. In this paper, we propose an STO detection and congestion window control algorithm based on the first acknowledgment following an RTO monitoring event for suppressing both the unnecessary retransmission and throughput degradation. This method strongly supports the enhancement of existing mobile communications systems because it does not require additional information or receiver modification.

抄録全体を表示

We propose a P2P file sharing system that enables flexible and intuitive information sharing and management among large group of users. The proposed system (NRBS: Network Resource Browsing System) supports a virtual directory that allows users to organize and manage distributed files based on simple and intuitive user interface. The system has a central management server that controls each user client in the system, which allows centralized security management and strict content control. In this paper, we compare conventional approach for managing P2P file sharing, and then propose a system based on virtual directory. We explain the mechanism for associating links on the virtual directory to actual file data stored in user clients' terminals. We evaluate the system based on usability, manageability, and security. The result shows that using virtual directory improves the usability and manageability while providing strict file security.

抄録全体を表示

In this paper, we propose several distributed zone partitioning schemes for Content-Addressable Networks (CAN), that is known as a pure peer-to-peer system based on the Distributed Hash Table (DHT). The main objective of the proposed schemes is to balance the load of nodes in the CAN system, in such a way that every node receives almost the same number of inquiries from the other nodes in the system. The result of simulations implies that, by using the proposed schemes instead of a randomized zone partitioning scheme originally implemented in the CAN system, we could reduce the response time for each inquiry to less than 75%.

抄録全体を表示

Our target is to support both small delay and small delay fluctuation of real-time traffic in IEEE802.11-based wireless LANs by using a decentralized manner. Several previous researches which aimed at supporting real-time traffic in IEEE802.11 wireless LANs developed decentralized control mechanisms achieving small delay of real-time traffic by differentiating real-time traffic from non-real-time traffic, but they cannot achieve small delay fluctuation because of the burst feature of the IEEE802.11 backoff mechanism. We propose a decentralized control mechanism for suppressing delay fluctuation in IEEE802.11-based wireless LANs. Our main proposal is a new backoff algorithm, called decentralized delay fluctuation control (DDFC), which can suppress delay fluctuation in a fully decentralized manner. DDFC can be easily used in IEEE802.11-based wireless LANs by replacing the current backoff algorithm of IEEE802.11 with DDFC. We examine the performance of DDFC, which is assumed to be used for real-time traffic in an IEEE802.11-based wireless LAN, by simulation. The results of computer simulation confirm that we can achieve not only small delay but also small delay fluctuation of real-time traffic in IEEE802.11-based wireless LANs by controlling real-time traffic according to DDFC.

抄録全体を表示

Today's Multi-player Online Games (MOGs) are challenged by infrastructure requirements because of their server-centric nature. Peer-to-peer overlay networks are an interesting alternative if they can implement the set of functions that are traditionally performed by centric game servers. In this paper, we propose a Zoned Federation Model (ZFM) to adapt MOGs to peer-to-peer overlay networks. We also introduce the concept of zone and zone owner to MOGs. A zone is some part of the whole game world, and a zone owner is a game sever of a specific zone. According to the demands of the game program, each node actively changes its role to a zone owner. By dividing the whole game world into several zones, workloads of the centric game server can be distributed to a federation of zones. In order to reduce response latency overhead on data exchanges between a zone owner and its clients, we limit the use of a Distributed Hash Table(DHT) to the rendezvous point of each zone; actual data exchanges are carried out through direct TCP connection between a zone owner and its members. We also use the DHT as backup storage media to cope with the resignation of a zone owner. We have implemented this zoned federation model as a middle layer between the game program and the DHT, and we evaluate our implementation with a prototypical multi-player game. Evaluation results indicate that our approach enables game creators to design scalable MOGs on the peer-to-peer environment with a short response latency which is acceptable for MOGs.

抄録全体を表示

The Future Vision Distribution Service (FVDS) (Sato, et al., 2004) is an innovative video delivery service concept based on the Internet ITS (Intelligent Transport System) area, and Source Mobility Support Multicasting (SMM) (Sato, et al., 2004) consists of new multicasting techniques needed to realize FVDS. However, FVDS and SMM lack sufficient application-level protocols to manage preceding and following vehicles on the same route, and have a drawback of topological constraints in the access network to accommodate micro-mobility. To solve these problems, this paper presents two new protocols. The Mobile Vehicle Management (MVM) protocol is an application-level protocol that discovers a preceding vehicle on the same route and provides its multicast address to following vehicles by using information from the global positioning system (GPS) and a route-guiding function. The Self-Organizing Tree (SOT) protocol is a network-level protocol for use in the access network that organizes radio base-stations into a logical tree topology in a self-forming and self-healing manner. This eliminates the topological constraints and provides the network with robustness against failures and flexibility for network design. To show how these protocols further the implementation of FVDS and SMM, this paper also gives details of a software implementation model for each network node system; specifically, we have designed a software architecture, composition elements for each protocol and their relations, and internal state-machines for each element.

抄録全体を表示

Today the WWW contains not only a tremendous amount of information but also a variety of Web applications, such as online shopping. Searching for such applications can be difficult and time consuming because current keyword-based approaches do not always deliver results that match the user's intention. The key underlying problem is that keywords do not capture the semantics of the user's query and the functional capabilities of Web applications. This paper presents a multi-faceted approach for searching Web applications. In our approach, the search process is based on various facets, such as service functionality, item (product), result type, inputs/outputs, and the detailed information of items. It is based on matching queries and Web application profiles that are described in DAML-S. The matching process is augmented with the use of ontologies. The result of applying our approach to a set of Web applications resulted in a precision of 93% and a recall of 99%.

抄録全体を表示

This paper proposes Proxy Certificate Trust List (PCTL) to efficiently record trusted delegation trace for grid computing. Our solution based on PCTL provides functions as follows: (1) On-demand inquiry service for real time delegation information of grid computing underway; (2) Lightweight mutual authentication that is beneficial for proxy nodes with short life span or limited computation power as wireless devices in mobile computing; (3) A kind of revocation mechanism for proxy certificates to improve the security and availability of grid computing.

抄録全体を表示

In this paper, we propose a new authentication mechanism for the mobile environments, called Self-Delegation. In the mechanism, a user stores information that relates to strict authentication in a tamper-resistant module that can be kept securely at home. Time-limited authority is delegated to the mobile terminal by communicating with the tamper-resistant module on a local basis. After the delegation, a remote service can authenticate the user for a limited time. We propose two self-delegation schemes, and analyze the security of the proposed scheme based on a security model that we define. Furthermore, we have implemented the self-delegation and authentication protocols on a PDA and a Java card, both of which have ISO14443 I/F, and show the feasibility of the implemented protocols.

抄録全体を表示

Verifiable secret sharing schemes proposed so far can only allow participants to verify whether their shares are correct or not. In this paper, we propose a new protocol which can allow participants not only to verify the correctness of their shares but also to revise the faulty shares. It is achieved in a cooperative way by participants, but without any assistance from the dealer. This protocol, to the best of our knowledge, is the first one providing such kind of ability. Correcting shares by participants instead of the dealer is important in many situations. In addition, this protocol is also useful for adding new participants without the dealer's assistance.

抄録全体を表示