National Institute of Informatics Open House Archives 2004

National Institute of Informatics
Open House 2004

* Archives 2003

Special Lecture

Title	Lecturer	Introduction
Ubiquitous Computing: Present and Future	SAKAMURA, Ken (Professor, University of Tokyo)	In the ubiquitous computing environment, the computers are embedded in everything around us, interconnected and collaborated through the networks, and support our lives behind the scenes. The ubiquitous computing (computer everywhere), which I have proposed and supported for a long time, as a new IT paradigm for the post-internet or the post-PC age, has been drawing enormous attention from many people. I will describe the technologies for realising the ubiquitous computing and their applications, and, furthermore, the social impact and future possibilities that it might bring.

<Introduction of Research Activities>

Picture


Research Division	Title	Lecturer
Foundation of Informatics Research	Logic as a foundation of informatics	TERUI, Kazushige
Infrastructure Systems Research	Traffic characterization and performance analysis using real traffic data	JI, Yusheng
Software Research	Approximate data matching for database integration	TAKASU, Atsuhiro
Multimedia Information Research	What can computers learn from TV broadcasts? - Knowledge discovery from large-scale broadcast video archives -	SATOH, Shin'ichi
Intelligent Systems Research	On Measuring and Analyzing Text-based Data	AIZAWA, Akiko
Human and Social Information Research	Extracting Translation Pairs from the Web	TSUJI, Keita
Research Information Research	Introduction of research case of Research Information Research field - A correlation of information related fields using keyword analysis -	NISHIZAWA, Masaki
Project / Development and Operations Department	NAREGI	USAMI, Hitohide
	International Scholarly Communication Initiative	Contents Division
	The Science Information Network ( SINET )/Super SINET	Network Division
	GeNii ( Global environment for Networked Intellectual Information )	Contents Division
	NACSIS-CAT/ ILL	Contents Division

<Introduction of the Ph.D.Program in Informatics>
HAYAMI, Ken

*Please click a title and you can see manuscript (PDF)

<Presentation and Exhibition>	Picture
Foundation of Informatics Research


Exhibition No.	Title	Author
101	Preconditioned GMRES Method for Least Squares Problems	HAYAMI, Ken , ITO, Tokushi
101	The prevalent method for solving large sparse least squares problems is to apply the Conjugate Gradient ( CG ) method to the normal equation , where A is a matrix. In this research, we propose a method of applying the Generalized Minimal Residual ( GMRES ) method to ABz=b or BAx=Bb where B is an appropriate nxm matrix. We give the necessary and sufficient condition such that this method gives a solution to the original least squares problem. Then, we give some numerical results validating the effectiveness of using an incomplete QR decomposition of A for B, compared to previous methods.	HAYAMI, Ken , ITO, Tokushi
102	The BioPortal	FUJIYAMA, Asao , TAKEDA, Hideaki , KITAMOTO, Asanobu , KAWAMOTO, Shoko , DEMIYA, Noboru, MIZUTA, Yoko , COLLIER, Nigel , SUGAWARA, Hideaki, ITOH, Takehiko, ARAKI, Jiro, YOSHINARI, Yasuhiko
102	Bio-portal is a 3-years project to construct portal site for various biology and biotechnology related data and databases. We are planning to develop user-freindly interface for students and teachers as well as researchers.
103	Speculative Computing	SATOH, Ken
103	We propose a method of distributed problem solving under incomplete information environment. To solve the problem, we propose a method using abduction. Abduction is a way of reasoning where some hypothesis will be used to complement unknown information. The idea is as follows. When communication is delayed or failed, then we use a default hypothesis as a tentative answer and continue computation. When some response is obtained, we check consistency of the response and the current computation. If the response is consistent, then we continue the current computation; else if the response is inconsistent, we seek an alternative computation. This way of computation is called speculative sincecomputation using a tentative answer would lead to a significant advantage if it succeeds.	SATOH, Ken
104	Collective spin decoherence in quantum computation	YAMAMOTO, Yoshihisa , MASTER, Cyrus, UTSUNOMIYA, Shoko
104	The information in quantum computers are often stored in identical spins ( or pseudo-spins ) that are separated by a distance shorter than the characteristic wavelength of a relevant to ensure a mutual cupling between two spins. Thus, the collective spin-reservoir interaction, rather than an individual spin-reservoir interaction, determines the decoherence characteristics. We use the symmetrized spin states to study this collective decoherence behavoirs in various quantum algorithms. The method of implementing quantum algorithms with stable quasi-dark state and avoiding unstable superradiant states and Schrodinger's cat states is proposed.	YAMAMOTO, Yoshihisa , MASTER, Cyrus, UTSUNOMIYA, Shoko
105	Finding Similar Substrings of Genomes	UNO, Takeaki
105	We propose an efficient algorithm for finding all similar parts of genome sequences. The problem has a very important roles in genome information science, such as diagnosing family diseases, finding genes, and constructing genealogical trees. We present the theory and the ideas of our algorithm, and demonstrate our implementation. Our algorithm can solve the problem more than 100 times faster than existing algorithms, and find exact solutions.	UNO, Takeaki
106	Rapid Content Annotation for Everybody on the Semantic Web	COLLIER, Nigel , KAWAZOE, Ai , MULLEN, Tony , WATTARUJEEKRIT, Tuangthong
106	The Portable Information Access ( PIA ) project aims to develop a domain adaptable information extraction ( IE ) system for annotating semantic content in texts. In contrast to other Web-based technologies such as information retrieval ( IR ) which are characterized by strong portability, no such system as yet exists for IE. We consider that the high cost of semantic annotation and its required expertise are potential bottlenecks in the spread of the Semantic Web. PIA is therefore concerned with machine learning for text-to-knowledge conversation so that computer programs can learn how to annotate new Web-based texts based on a relatively small number of examples of annotated texts in the domain. We are now focussing on machine lelearning ( supportector machines, maximum entropy, hidden Markov models etc. ) for the major tasks of named entity, coreference resolution and predicate-argument structure annotation and are conducting large-scale experiments in general news and molecular biology domains.
107	Zone Analysis in Scientific Articles	COLLIER, Nigel , MIZUTA, Yoko
107	In this project we have begun this year an investigation into the automatic identification of regions corresponding to rhetorical types in scientific and technical articles. For example where the author mentions his/her specific aim, background publications, specific results of his/her experiments, explains the significance of the results, contrasts the results to previous work etc. This will be highly useful in information extraction ( i.e. fact extraction from text ) as it reduces the amount of text that needs to be searched in order to locate possible facts. It also has applications in other areas such as summarization of documents where the summery needs to be generated from only certain zones in the document. The key point of our approach is to do this using machine learning from a set of pre-annotated articles using maximum entropy ( ME ) and support vector machines ( SVMs ).	COLLIER, Nigel , MIZUTA, Yoko
108	An ecological approach to coordination among speech, gesture and breathing movements in natural conversation	FURUYAMA, Nobuhiro , TAKASE, Hiroki , HAYASHI, Koji
108	This poster reports a study on coordination among speech, gesture and breathing movements in quasi natural conversation and compare the result with a series of studies we have conducted on their coordination in more controlled experimental settings.	FURUYAMA, Nobuhiro , TAKASE, Hiroki , HAYASHI, Koji
109	Interactive Proof Systems in Quantum Computation and its Applications	MATSUMOTO, Keiji
109	Interactive proof system and related computational models play very important role in contemporary computational complexity theory. Here, quantum versions of those are reviewed. Application to zero-knowledge proof is also explained.	MATSUMOTO, Keiji
110	Quantum coherence and quantum information processing	NEMOTO, Kae
110	The output from a Laser has been considered coherent light, and hence used for quantum information processing. However, it has recently been argued that the inability to measure the absolute phase of the electromagnetic field prohibits the representation of a laser's output as a quantum optical coherent state. In this presentation, we establish a fundamental principle to represent a quantum system when the system involves quantities which in principle cannot be measured. Furthermore we introduce new proposals to implement essential operations for quantum computation in optics.	NEMOTO, Kae
111	Computational Complexity and Proof Complexity	TERUI, Kazushige
111	Computer programs are often verified by mathematical proofs, while mathematical proofs are often given by demonstrating programs ( algorithms ). Thus it is naturally expected that there is some deep connection between programs and proofs. Here we study this connection from the viewpoint of "complexity". We propose several measures for proof complexity P and other lower computational complexity classes.	TERUI, Kazushige
112	Ontological Knowledge Base Reasoning with Sort-Hierarchy and Rigidity	KANEIWA, Ken , MIZOGUCHI, Riichiro(Osaka University)
112	We present an extension of an order-sorted logic with ontological property classification. This logic types ( rigid sorts ), non-rigid sorts and unary predicates to distinguishably express the properties: substantial sorts, non-substantial sorts and non-sortal properties. In this study, we propose a reasoning algorithm for many separated knowledge bases where each knowledge base can extract rigid property information from other knowledge bases.	KANEIWA, Ken , MIZOGUCHI, Riichiro(Osaka University)
113	Development of next generation human interface	UEKI, Kouichirou
113	Next generation human interface to improve human-computer interaction is inevitable for high performance computing environment. Genetic algorithm for solving traveling salesman problem was proposed to handle high dimensional information such as Internet transactions. The new genetic algorithm will enable us to realize the next generation computing environment with the high level human interface.	UEKI, Kouichirou
114	Quantum Information Processing by Linear Optics	van Loock, Prter
114	An essential tool in quantum information processing ( quantum computing nd quantum communication ) is to perform quantum state measurements. When the quantum information is encoded into optical states, the most practical way to implement such measurements is by means of linear optics using only beam splitters and, for example, photon counting. We discuss how one can easily assess whether a given measurement can be indeed implemented via a feasible linear-optics scheme.	van Loock, Prter

Infrastructure Systems Research


Exhibition No.	Title	Author
201	Field Test for Establishment of Next Generation Optical Networks	ASANO, Shoichiro , FUJINO, Takayuki , YONEDA, Susumu
201	In order to establish "real" next-generation optical networks, various network features such as transmission fiber features or optical path rerouting status transision features are need to be investigated. This presentation mainly consists of the one presented in OFC2004, and treats field test results of 40Gbps transmission on the transmission fibers and result of field of GMPLS all-optical path rerouting.	ASANO, Shoichiro , FUJINO, Takayuki , YONEDA, Susumu
202	Multi-layer Integration Control Architecture for Next-Generation Optical Networks	ASANO, Shoichiro , FUJINO, Takayuki , URUSHIDANI, Shigeo
202	Network today is designed to use "layer model". This means individual layer is managed and operated separately. But next-generation optical network will be able to user-oriented lambda path deployment. For example, if certain user application signals to their demand of lambda. then network respond and provides lambda path for the application. This requires that multiple layer work and cooperate together at a time. This presentation introduces "multi-layer integration control architecture", which can manage and operate multiple layer at a time using GMPLS scheme.	ASANO, Shoichiro , FUJINO, Takayuki , URUSHIDANI, Shigeo
203	A New Privacy Protection Technology in Ubiquitous Computing Networks	YAMADA, Shigeki
203	A new privacy protection technology is proposed that handles privacy-sensitive information such as personal data without causing privacy invasion in ubiquitous computing networks.	YAMADA, Shigeki
204	A high-level synthesis tool for asynchronous circuits	YONEDA, Tomohiro
204	Asynchronous circuits with no global clock systems have potential for solving several problems that are facing current computer systems, such as clock skew and high power consumption. This research is developing an automated synthesis tool for asynchronous circuits, which takes a high-level specification language SpecC, performs resource scheduling/allocation, and generates gate-level circuits.	YONEDA, Tomohiro
205	Network Control Scheme for Very Fast Internet	MATSUKATA, Tomohiro
205	Due to the increasing transmission speed of the Internet and growing power of computer systems, it is very likely that very fast communication between computers that we have not ever expected is emerging. It is urgent to establish the network control scheme for providing such fast communication in a stable manner. Not only layer 3 technologies including MPLS, traffic engineering, and VPN, which control Internet, but also technologies like GMPLS which interworks with lower layers are important.	MATSUKATA, Tomohiro
206	Providing the Quality of Service in Multi-Service Networks	JI, Yusheng
206	Providing the quality of service based on the requirements of applications of users is necessary in networks shared by all kinds of users and applications. This research is focused on traffic analysis and control methods for QoS guarantee in multi-service networks.	JI, Yusheng
207	SSS-PC: The next generation operating system	MATSUMOTO, Takashi
207	Takashi Matsumoto and his project members are developing a new generation operating system named SSS-PC ( three S PC ) which runs on commodity PCs in offices and makes up a highly-dpendable and high-performance computer system. SSS-PC can integrate up to a hundred thousands of PCs into a single high performance parallel computer. Furthermore, SSS-PC lets us ers to maintenance works such as machine replacement or component inspection without stopping running applications, and dynamically extend or reduce system configuration without system suspension. LINUX or UNIX applications can be ported to SSS-PC without modifying source codes by using compatible program development environment. Popularity is planned to be gained from server computers where high availability and non stop requirements for applications are high.	MATSUMOTO, Takashi
208	QoS evaluation method and network dimensioning tool for the Internet	ABE, Shunji , HASEGAWA, Toru
208	To realize QoS control and the bandwidth dimensioning which satisfies a required QoS, it is necessary to estimate QoS from the Internet traffic flow. Many traffic-patterns with the different propertties ( Short-Range Dependence, Long-Range Dependence ) of an autocorrelation are flowing in the Internet. The properties make their influence on the QoS very different. As one of QoS estimation methods, we show an approach that estimates the QoS by selecting one out of the two properties according to the offered traffic intensity in the Internet, and also demonstrate a network dimensioning tool by our approach.	ABE, Shunji , HASEGAWA, Toru
209	Study on Context-Aware Communication Networks	KAMIOKA, Eiji
209	Context aware technologies are mandatory for the realization of communication services in a ubiquitous environment. The significance of such technologies is how the communication service system handles users' context and provides the content or the communication service for the user. We will present the concepts of context aware communication networks and demonstrate a prototype system based on some communication service scenarios.	KAMIOKA, Eiji

Software Research


Exhibition No.	Title	Author
301	Research on an extensible distributed Operating System	MARUYAMA, Katsumi , KODAMA, Kazuya , HIDAKA, Soichiro , HASHIZUME, Hiromichi , JI, Yusheng , OZAKI, Ryouta
301	A very large amount of control software is being developed every year, and it is highly desired to improve program development productivity. ( Here, control software is a programs used in control systems, embedded systems etc.) Control systems are required to provide very specific functions, such as sophisticated distributed processing. To answer these requirements, an approproate Operating System is desired; general purpose OS's are not sufficient, because requirements differ from system to system and OS extension is very difficult. In this research, a new Operating System structure has been studied. It is based on L4 micro-kernel and multi-server scheme. 2 types of implementation ( L4minix-1, L4minix-2: free source ) are now start working.
302	Next Generation Software for Mobile and Ubiquitous Computing	SATOH, Ichiro
302	I research a middleware system for managing distributed systems as a next generation software technology for mobile and ubiquitous computing. The middleware can provide adaptive, reliable, self-organized management for such systems. An implementation of the system and its demonstration are presented.	SATOH, Ichiro
303	Constraint Programming and its Application to User Interfaces	HOSOBE,Hiroshi
303	Constraint programming, which allows developing programs by declaratively specifying constraints, is a powerful problem solving method used in fields such as artificial intelligence, logic programming, and user interfaces. We present constraint satisfaction, which is fundamental to this method, and give constraint-based user interface construction, focusing on recent results of our research.	HOSOBE,Hiroshi
304	'mono project': Multimedia Archive for Creative Activities of Arts and Crafts	AIHARA, Kenro , TAKASU, Atsuhiro , FUJIHATA, Masaki, KATSURA, Eishi, YOKOMIZO, Hiroko, ITOH, Yoshiaki
304	'mono project', collaboration with Tokyo Nat. Univ. of Fine Arts and Music and Iwate Prefectural Univ., focuses on research issues related to developing and using multimedia archive including text, image, video, and audio, about creative activities of arts and crafts. This presentation overviews our project and shows one approach of interactive system.
305	Optimizing XQuery using Formal Semantics	HIDAKA, Soichiro , KATO, Hiroyuki
305	XQuery is a functional query language for XML. It is defined in terms of formal semantics. We try to exploit this functional aspect of the language and apply some traditional optimization techniques such as common subexpression elimination and constant folding. XQuery-specific aspects of the optimizations are also studied. Lambda calculus is used for discussing equalities between expressions before and after transformation.	HIDAKA, Soichiro , KATO, Hiroyuki

Multimedia Information Research


Exhibition No.	Title	Author
401	Parametron and its applications	HASHIZUME, Hiromichi
401	Parametron is a logic device invented Japan in 1950s. Besides its unique constraction and operation, currently nobody try to build it as there is no supply of special magnetic material for parametrons. In this presentation I will remake the device and introduce its applications.	HASHIZUME, Hiromichi
402	Construction of Tokyo Bunko Valuable Books Multimedia Digital Archive	YAMAMOTO, Takeo , KITAMOTO, Asanobu , SATO, Sonoko , ONO, Kinji
402	Valuable books held by Tokyo Bunko, which is one of the five major oriental study libraries in the world, was organized into a multimedia database and made multilingually accessible from the Web. Technologies used in organizing the digital archive, including Multilingual Annotation Support System ( MASS ), are described.
403	Web page clustering on parameterized connectivity	ADACHI, Jun , TAKASU, Atsuhiro , MASADA, Tomonari
403	WWW includes many meaningful information units as Web page clusters. Hence we divise a new Web page clustering method. We define the distance between two pages as page weight sum along the cycle connecting them, and we provide a parameter for limiting cluster member enumeration range as cluster size control tool. Our clustering is a generalization of SCC decomposition. We also show experimental results.	ADACHI, Jun , TAKASU, Atsuhiro , MASADA, Tomonari
404	A study on structuring large-scale broadcast video archives	SATOH, Shin'ichi , KATAYAMA, Norio , MO, Hiroshi , KATO, Hiroyuki
404	We are investigating technologies to delineate necessary information upon users' requests from large-scale broadcast video archives combining image analysis, natural language processing, data mining, information retrieval, artificial intelligence, etc. Our primary goal is to establish a method to somehow understand the contents of broadcast video streams by computers.
405	Image Generation Using Transformation Integrating Multiple Differently Focused Images	KODAMA, Kazuya
405	A method of image generation using transformation integrating multiple differently focused images is shown. First, acquired images with a certain condition and some 3-D information of objects are combined by using a 3-D blurring transformation. Then, based on a deconvolution using 3-D FFT, for example, and all-focused image can be reconstructed.	KODAMA, Kazuya
406	The ASPICO platform	ANDRES, Frederic
406	This research proposes the Advanced Scientific Portal for International COoperations on DSR （APICO-DSR ）, methods and tools to understand, to manipulate, and to share the free-copyright Digital Silk Roads resources and to relevant set up and high performance distributed services.	ANDRES, Frederic
407	A Multi-lingual Multimedia Ontology for Caravanserai Management	ANDAROODI, Elham, ANDRES, Frederic , ONO, Kinji , COLLIER, Nigel , LEIBIGRE, Pierre
407	I will introduce the research regarding to the development of a multi-lingual multimedia ontology describing different physical and spatial components of caravanserai architectural object.
408	3D Shape Comparison using Multiview Images	GOTODA, Hironobu
408	Comparing and matching 3D shapes is a fundamental problem in 3D shape modeling, retrieval, and recognition. We address this problem using the similarities among the images associated with multiple viewpoints. Our method can be applied to many types of shape representations, including polygonal models, voxels, and image-based representations.	GOTODA, Hironobu
409	Optimizing XQuery by "FLWOR Arranging"	KATO, Hiroyuki , HIDAKA, Soichiro
409	XQuery can be mapped to the queries of the underlying models, and can use their optimization features. However some information sources such as file systems may not support optimization functionality. Therefore developing optimization techniques by rewriting XQuery is very important. We propose a rewriting technique called FLWOR Arranging for optimizing XQuery with nested FLWOR blocks.	KATO, Hiroyuki , HIDAKA, Soichiro
410	X-informatics: Informatics in the Age of Large-Scale Database Technology	KITAMOTO, Asanobu
410	What is X-informatics? What is new, and how informatics will change from the idea? The author summarizes these points with the result of projects that the author is involved in for these past few years with the aim of building large-scale image database portals on earth science data and cultural heritage data. The author also introduces research results on database navigation/ retrieval and data mining worked on those real-world databases.	KITAMOTO, Asanobu

Intelligent Systems Research


Exhibition No.	Title	Author
501	Development of Distance Education System WebLS	UENO, Haruki , AMPORNARAMVETH, Vuthichai , ZHANG, Tao , SAE-KOK, Worawut(RA) , MORI, Masaki, ISHIDA, Masayuki(Ohmsha Ltd.)
501	Web-based learning is becoming an important tool in education. However, the cost of producing an on-line course is high. WebLS system is aiming to assist traditional instructors to archive their learning materials on the web. WebLS features web-based authoring tools, portability on various platforms, and flexibility of XML-based storage allowing the learning materials to be delivered in many different ways.
502	Next-Generation Symbiotic Robot System	UENO, Haruki , ZHANG, Tao , AMPORNARAMVETH, Vuthichai , PATTARA, Kiatiseve , MD, Hasanuzzaman, SHIRAI, Yoshiaki , MATSUI, Tomoko
502	As a next generation robot system, symbiotic human-robot system will play an important role in the future welfare society. In order to construct a high intelligent, human-friendly symbiotic human-robot system, a knowledge-based dialogue manager and PCA based gesture recognition are developed for human-robot interface. Cooperative operation of robots using Software Platform for Agents and Knowledge Management ( SPAK ) is proposed to perform complex tasks of symbiotic human-robot system.
503	Construction of Gourmet Advisor Using Concept of Complex Adaptive Information System	UENO, Haruki , AMPORNARAMVETH, Vuthichai , ZHANG, Tao , KOVACS, Alexander , KUROMIYA, Shizuka, ISODA, Yoshinori, KURAKAKE, Soji(NTT DoCoMo)
503	iGuruAdo is an information system to find restaurants. It differs from common such systems in that it is based on a new architecture in which pieces of information and their interaction are of prime interest. I call these systems complex adaptive information system; they are comprised of a substrate and workspaces. The substrate contains an evolving meme network. Interaction of memes happens in the workspaces, such that there is usually on workspace for one user.
504	Design of Interaction Embedded in a Task for Cooperative Sweeping	YAMADA, Seiji , KOBAYASHI, Kazuki
504	We propose Interaction Embedded in a Task for human-robot interaction and apply it to cooperative sweeping by a human and a miniature mobile robot. The demonstration is presented.	YAMADA, Seiji , KOBAYASHI, Kazuki
505	Knowledge Symbiosis	TAKEDA, Hideaki , ICHISE, Ryutaro , HAMASAKI, Masahiro , OHMUKAI, Ikki , NUMA, Kosuke
505	The aim of this research is to propose an architecture to enable people with different knowledge to exchange and distribute their knowledge. For example, we developed blog-based systems to support human information and communication activities seamlessly.
506	On Measuring and Analyzing Web Texts	AIZAWA, Akiko
506	This presentation introduces our recent attempt on text targeted to Web. In our research, we investigate term sequence distributions shared by documents within/ between Web sites using large-scale Web collections. We further utilize the result in such applications as key phrases extraction, Web-specific expression ,identification and related Web sitesdetection.	AIZAWA, Akiko
507	"Active contents"- Contents distribution by mobile agents -	HONIDEN, Shinichi , TAHARA, Yasuyuki , YOSHIOKA, Nobukazu , students of Honiden lab.
507	Management of contents which are distributed on various networks becomes increasingly important for contents providers and users. We adopt mobile agent technology to treat the contents pollution. This research is supported by Strategic Information and Communications R&D Promotion Scheme in the Ministry of Public Management, Home Affairs, Posts and Telecommunications.
508	Discovery of Communities in the Web	MURATA, Tsuyoshi
508	Finding related pages from huge Web network is important for assisting users' information retrieval from the Web. In this presentation, we will explain a method for discovering and visualizing Web communities ( groups of Web pages sharing common interests ) based on connectivity information of hyperlinks. The method is based on the assumption that hyperlinks to related Web pages often co-occur. A system based on the method succeeds in discovering several genres of Web communities without analyzing the contents of Web pages.	MURATA, Tsuyoshi
509	Reflectance Estimation from Motion under Complex Illumination	FEI, Do , OKABE, Takahiro , SATO, Yoichi , SUGIMOTO, Akihiro
509	We propose a method for recovering the reflectance properties of a convex Lambertian object from and image sequence of the object in motion under unknown, complex illumination. Our proposed method combines the geometry reconstructed by shape-from-motion with the spherical-harmonic representation of the Lambertian reflectance under arbitrary illumination, and, then, recovers the albedo of the object and the illumination distribution. The proposed method enables us to synthesize realistic images of the object in arbitrary poses under arbitrary lighting conditions. We conducted a number of experiments by using both synthetic and real images and confirmed the effectiveness of the proposed method.
510	Chemoinfomatics Research - Practical Use of Chemical Information and Computer for Chemistry Research	SATOH, Hiroko
510	Our chemoinfomatics research which is aiming at intelligent systems for solving practical problems in chemistry is briefly demonstrated. It contains effective utilization of both computer and chemical information, the creation of new representation of chemical information which is suitable for human and computer, respectively, the proposal of advanced chemical information contents, and the development of some intelligent systems for chemistry. Computer systems for chemical reaction prediction and NMR ( Nuclear Magnetic Resonance ) spectrum prediction are shown in the presentation.	SATOH, Hiroko
511	Knowledge Discovery in Medical Databases	ICHISE, Ryutaro
511	Because information systems are used in large hospitals, a large amount of medical data is provided to physicians. This large amount often constitutes an information overload. Consequently, computers must extract useful information from such data. The most difficult issue in managing medical data is the handling of time-series data, which contain irregularities. In this presentation, we describe handling time-series data using a first-order logic discovery algorithm. In order to evaluate our method, we conducted an experiment using real-world medical data. The results show that our algorithm is able to induce knowledge involving a temporal relationship.	ICHISE, Ryutaro

Human and Social Information Research


Exhibition No.	Title	Author
601	Finding Domain Term Structure based on Verb Collection	KOYAMA, Teruo , KAGEURA, Kyo
601	In this presentation, we discuss about a method for finding term structure in a domain text corpus, based on the collocation of terms and verbs within documents. We also show results of a trial to classify terms according to the revealed structure.	KOYAMA, Teruo , KAGEURA, Kyo
602	Utilizing anchor information and link information for searching known items on the web	OYAMA, Keizo, 江EGUCHI, KOji , AIZAWA, Akiko , ISHIKAWA, Haruko
602	Based on the results of Navigational Retrieval Subtask conducted in the NTCIR-4 WEB Task, the authors discuss, from various aspects, the effects of using anchor information and link information for known item search, i.e., searching a representative page of a given known item, and show their effective usage.
603	Value-added information Access Systems	KANDO, Noriko , EGUCHI, Koji , SEKI, Yohei , NAKAYAMA, Norio , KAJIYAMA, Tomoko
603	This exhibition introduces the outline and examples of the value-added information access systems. The information access technologies like information retrieval text summarization have been focused on "topic" or "aboutness" of the documents, so far. However, it has been known that the context like user's situation, tasks, and of access, plays critical roles in the access and user's satisfaction. Multi-faceted flexible access to the information is required. This exhibition overviews the strategies and introduces some examples including; Multi-Document Summarization Based on User's Viewpoints, Emotion Identification from Text, GUI for Book Search System, and Concentric Ring View: An Interactive Environment for Integrating Multi-Faceted Search and Browsing.
604	Information Institution Research; Building a Weblog Site for Online Law Information	TOHKURA, Yoh'ichi , OKADA, Hitoshi
604	There is great gap between the technological reality and the legal assumption in this information-oriented society. In order to remove this gap as possible, Information Institution Research Division researches: the interpretation of the regulations now in force, the clarification of legislation problem to be solved, and the enactment of information network law. We are preparing a weblog site as a space of opinion gathering, discussion, and agreement to catch up the sped of change, and end of which is introduced this time.	TOHKURA, Yoh'ichi , OKADA, Hitoshi
605	Detecting bilingual paraphrase/ variation patterns of terms	KAGEURA, Kyo , YOSHIKANE, Fuyuki , NOZAWA, Takayuki
605	We have developed the parallelised ruled of term variations in Japanese and English, and implemented the rules into a system which detect actual variations from the web.	KAGEURA, Kyo , YOSHIKANE, Fuyuki , NOZAWA, Takayuki
606	Online clustering for Web information access and their evaluation methods	EGUCHI, Koji
606	The author reports on online clustering methods that classifies retrieved documents on-the-fly, and evaluation methods for such techniques. The evaluation was performed from various points of view using retrieved documents obtained by several types of Web search engines.	EGUCHI, Koji
607	Predicting the Lifespan of Newly-coined Terms	TSUJI, Keita
607	We present a method for recognizing the terms which are to prevail in special domains. The data for investigation is a set of academic papers published during these 20 years and the time-serial trend of the terminology before the terms appeared is used as a key clue for forecasting. Such method will help the lexicographer to update the dictionaries and give useful information to trend analysis.	TSUJI, Keita

Research Information Research


Exhibition No.	Title	Author
701	Research on citation impact of papers in Japanese academic journals based on NII CJP database	NEGISHI, Masamitsu , Sun, Yuan
701	In recent development of Japanese universities' reformation, research evaluation has rapidly become a hot topic. By using citation index databases, number of publications and their citations can be calculated, which could be used as an index of research level. As the Impact Factors derived from Science Citation Index database by ISI is especially getting popular, their improper use is often pointed out. NII is constructing a Citation Database for Japanese Papers ( CJP ) with a wide coverage of Japanese academic journals both in Japanese and English. Some of our bibliometric analyses based on CJP will be shown.	NEGISHI, Masamitsu , Sun, Yuan
702	A correlation of information related fields using keyword analysis	NISHIZAWA, Masaki , Sun, Yuan
702	In a lot of cases, current classification scheme does not work well for classifying new research fields into actual disciplinary structure. This is especially significant for information science. In response to such a situation, application fields of Grant-Aid for Scientific Research were greatly revised especially in the multi-disciplinary region in 2003 fiscal year. Here, we report about the correlation of the old and the new classification in the information related fields using the adopted research subject and keyword.	NISHIZAWA, Masaki , Sun, Yuan
703	Research on Creativity in R & D Activities	SHIBAYAMA, Morio , YANO, Masaharu
703	In order to promote scientific research, research proposals with creativity would early be distinguished and supported, and environment bearing originality should be made clear and developed much more. Themes of research are analysis on environment cultivating creativity, study on concepts of creativity and identification of creative researches, and evavuation of creativity in the R & D activities od universities and corporate organizations. We introduce the topics of these researches.	SHIBAYAMA, Morio , YANO, Masaharu
704	Research and Innovation Policies in France and Japan: similarities and differences	ANGELINO Henri , COLLIER, Nigel
704	During the last 20 years many change occur concerning research and innovation policies either in Japan or in France. In France it corresponds to the "disappearance of the Colbertiste State" which culminates with the Innovation Act in 1999 and later on the incubators policy's implementation, actions for facilitating seed funds, relation university-industry, etc. And some more recent Laws passed in 2003. In Japan Science and Technology has always been a priority but the Government adopted a new approach in the 1990's. The starting point was the S&T Basic Law in 1995 and since then many actions have been taken and Laws enacted：S&T Basic Plan, TLO creation, Japanese Bayh-Dole Act, creation of the Council for Science and Technology Policy, Intellectual property Law, etc. Both countries are "moving in the same direction", using the "same recipes" though adapting the solutions to their national context.	ANGELINO Henri , COLLIER, Nigel
705	Bibliographic utility: it's trend and prospect	MIYAZAWA, Akira
705	Bibliographic utilities, such as NACSIS-CAT/ ILL, are being one of the infra structure in national level information service, together with the national library. This presentation introduces bibliographic utilities of the world, North America, Europe and East Asia. And, recent trends of OCLC and RLIN in US, KERIS in Korea and CALIS in China are analyzed from technical and management aspects with future prospects.	MIYAZAWA, Akira
706	Misunderstanding and trap of information sharing	KOHARA, Masahiro
706	In the information-oriented society, some kind of information, which is even unclear and changeable, is often somehow shared by people in the contexts of its historical development, its superiority, its relativity, its image, and so on. It is probably the desirable attitude in information age that we overcome the dependency to the word with a trump effect as a information sender and the easy acceptance of it as the receiver. From this viewpoint, I would like to clarify the traps which information sharing tend to fall into with quoting some examples.	KOHARA, Masahiro
707	An Analysis on factors in the difference of e-learning diffusion in Japanese and American Higher Education	WATANABE, Keiko
707	Japanese universities have been actively discussing introduction of an e-learning as a part of their educational reform. This analysis aims to verify the factors that cause differences between USA and Japan of how and to what extent such e-learning has been adopted. The result of this analysis can provide clues to effective introduction of an e-learning to Japanese universities.	WATANABE, Keiko

Graduate Students


Exhibition No.	Title	Author
801	Difficulty of Graph Isomorphism	KIYOMI, Masashi , UNO, Takeaki
801	Given two graphs, graph isomorphism is to determine whether the two graphs have same structure. It has been open problem for long time how difficult graph isomorphism is, and it is regarded as a problem between easy problems and difficult problems. In our presentation, we want to enjoy the difficulty of graph isomorphism with our puzzle program.	KIYOMI, Masashi , UNO, Takeaki
802	Supporting Quality-of-Service ( QoS ) on Optical Internet	PHURITATKUL, Jumpot , JI, Yusheng
802	Optical Burst Switching ( OBS ) is a promising to support the next generation optical Internet. In order to provide service differentiation in OBS networks, we propose a buffer allocation algorithm to schedule burst at edge OBS nodes and a bandwidth allocation algorithm and a Fiber Delay Line ( FDL ) allocation algorithm to schedule burst at core OBS nodes. We also introduce a new burst assembly technique. The controllable burst loss probability for different service classes is achieved by the proposed algorithms.	PHURITATKUL, Jumpot , JI, Yusheng
803	A study on fast handover protocols in Mobile IPv6	LI, Lei , ABE, Shunji
803	Many micro-mobility protocols have been proposed to provide fast handover. We discuss the defects of some exiting micro-mobility schemes, propose a new Xcast based scheme, analysis the improvement of handover performance of the proposed scheme and then do the simulation to compare the different handover performance of HMIPv6, Multicast based HMIPv6 and the proposed Xcast based HMIPv6 scheme. The simulation results show that our proposal is an efficient and scalable scheme to achieve fast handover with less signaling overhead.	LI, Lei , ABE, Shunji
804	A burst multiplexing method by use of selective burst acknowledgement mechanism for the optical switching network	DU, Ping , ABE, Shunji
804	Optical Burst Switching ( OBS ) has been expected as one of the switching technologies for the next generation optical transport network. On realizing an efficient burst multiplexing, the assembly granularity and the offset time between a control packet and its corresponding payload are two important parameters of OBS network. They vary based on the circumstance of a network and affect the network efficiently. However, they are beyond study now. In this presentation, we will show a solution for the optimal parameters by your proposed mechanism - Selective Burst Acknowledgment Mechanism.	DU, Ping , ABE, Shunji
805	Image Processing Motivated Structures and Functions of The Visual Cortex - Detection of moving objects their motions and colors -	NAKAJIMA, Chikahito , UENO, Haruki
805	The Middle Temporal area ( MT ) and the Visual area 4( V4 ) are connected in the visual cortex in the brain. MT cells are specialized for detection of movement and V4 cells are specialized for color. We refer to the structures and functions of MT/V4 and propose a detection method of objects using their motions and colors. We also describe that another visual area is required to detect moving objects from the point of view of computational vision.	NAKAJIMA, Chikahito , UENO, Haruki

Project/Development and Operations Department


Exhibition No.	Title	Division/Author
A01	The Science Information Network ( SINET ) / Super SINET	Network Devision
A01	The Science Information Network ( SINET ) is an information communication network which connects universities and academic research institutions in Japan. Super SINET is an ultrahigh-speed network which is intended to promote advanced academic researches in Japan. The e-Japan Priority Policy Program announced by the IT Strategic Headquarters in March 2001 referred to this network, the world's fastest Internet for research, based on 10 Gbps optical communication technology.	Network Devision
A02	NAREGI	The Collaborative Center for Research Grid
A02	NAREGI ( National Research Grid Initiative ) is one of the collaboration projects among industry, academia and government, initiated by the Ministry of Education, Sports, Culture, Science and Technology ( MEXT ). At the Center for Grid Research & Development, our research and development area are the grid middleware and the networking technologies; the Grid is regarded as the foundation for the information technology in the 21st century. Japan's leading institutions from industry, academia and government also take part in NAREGI Project, and it is expected that the fruits of the project will boost the research and development in the related scientific fields, and furthermore, will lead to the strengthening of Japan's competitiveness in the world's economy.	The Collaborative Center for Research Grid
A03	Studies on Integrating Scholarly Information	Research Center for Testbeds and Prototyping
A03	NII has various scholarly databases and is now carrying out a project to integrate them for providing users with an integrated and uniform access to them. In this presentation, we report integrated and presentation methods for bibliographic and researchers' information in various databases.	Research Center for Testbeds and Prototyping
A03-1	Development of a retrieval method considering matching errors in integrated databases	OYAMA, Keizo , AIZAWA, Akiko , GOTODA, Hironobu
A03-1	The authors report on method that realizes an efficient and appropriate retrieval function for integrated bibliographic databases where matching uncertainty are included by utilizing confidence of matching.	OYAMA, Keizo , AIZAWA, Akiko , GOTODA, Hironobu
A03-2	Linkage of an editing system for researchers' personal data to CiNii	OYAMA, Keizo , ADACHI, Jun , AIZAWA, Akiko , KATO, Hiroyuki , IDE, Ichiro
A03-2	The system was developed for enabling uniform and efficient publication of researchers' information at universities and academic institutions. Its functions for linkage with CiNii are reported.
A03-3	A Fast Method for Duplicated Entries Detection in Bibliographic Databases	AIZAWA, Akiko , OYAMA, Keizo , TAKASU, Atsuhiro , ADACHI, Jun
A03-3	We propose a method for identifying duplicated entries in bibliographic databases, and show some experimental results using real-scale databases that are serviced in GeNii.
A03-4	Bibliographic String Parser and Bibliographic Component Extraction	TAKASU, Atsuhiro , ADACHI, Jun
A03-4	We report a statistical model for parsing bibliographic strings and extracting bibliographic components. The reported model is applied to articles in NII ELS for citation linkage.	TAKASU, Atsuhiro , ADACHI, Jun
A03-5	Visualization of information on researchers - a meet-a-mate site for researchers -	TAKEDA, Hideaki , ICHISE, Ryutaro
A03-5	Activities of researchers are supported by various communities which are formed by research interest and human relationship. We show a prototype system which can visualize researchers information and communities by calculating relationship based on CiNii database.	TAKEDA, Hideaki , ICHISE, Ryutaro
A03-6	GeNii ( Global environment for Networked Intellectural Information )	GOTODA, Hironobu , MIYAZAWA, Akira , OYAMA, Keizo , AIZAWA, Akiko , KANDO, Noriko , SUN, Yuan , NISHIZAWA, Masaki , MURATA, Tsuyoshi、 HOSOBE, Hiroshi
A03-6	This presentation reports on the prototype implementation of an add-on to CiNii ( NII citation information navigator ). This add-on enables us to visualize citation relationships among published papers, providing an easy-to-use interface for navigating though the repository of journals.
A04	NACSIS-CAT / ILL	Contents Division
A04	GeNii ( Global environment for Networked Intellectual Information ) is building an environment that can facilitate the integrated use of information needed for scientific research by linking contents provided through the various NII services and other valuable intellectual information resources in Japan and overseas spotted on the Internet. At present "NII Article Information Navigator", "Webcat Plus ( Book Information Navigator by NII )," "Information of Research Topics and Research Results" and "Portal to Japanese Universities" ( JuNii ) are prepared aiming at the formal operation in April, 2005.	Contents Division
A05	International Scholarly Communication Initiative	Contents Division
A05	NACSIS-CAT/ ILL is service for libraries to construct union catalog database of books and serials in the libraries and to provide the books and article of serials for all researchers and students in Japan. The union catalog database is offered to public through Webcat and Webcat Plus. At present, there are over 70 million books are registered.	Contents Division
A06	NTCIR: Evaluation Workshop for Information Access Technologies	Contents Division
A06	International Scholarly Communication Initiative is a project that is encouraging formation of an electronic journal and promoting internationalization of English academic journal in Japan in collaboration with university libraries and Japan Science and Technology ( JST ). At present, we are supporting 21 titles of English paper over 16 academic societies to create business model, such as site license together with university libraries.	Contents Division
A07	Cross-lingual Information Retrieval Task	Research Center for Informaton Resources (KANDO, Noriko , ADACHI, Jun)
A07	The exhibition introduces the NTCIR ( NII Test Collection for IR Systems ) workshop, a series of evaluation workshops on information access technologies like information retrieval, question answering, and summarization, aiming to encourage the research on information access technologies by providing a large-scale infrastructure like reusable test collections and the forum of researchers. In this exhibition we provide the project overview and introduce some examples of the participating systems using the NTCIR test collections. The 4th NTCIR Workshop meeting will be held on June 2-4, 2004, right after the NII Openhouse and everyone is welcome to join.
A07-1	Overview of Patent Retrieval Task at NTCIR-4	KISHIDA, Kazuaki , KURIYAMA, Kazuko(Shirayuri College), KANDO, Noriko , EGUCHI, Koji
A07-1	Cross-Lingual Information Retrieval ( CLIR ) is the circumstance in which a user tries to search a set of documents written in one for a query in another language. In this project, CLIR on East-Asian Languages ( Chinese, Japanese and Korean ) and English is experimentally investigated.
A07-2	An Evaluation of Question Answering Challenge 2	FUJII, Atsushi(University of Tsukuba), IWAYAMA, Makoto(Tokyo Institute of Technology/Hitachi Ltd.), KANDO, Noriko
A07-2	We performed the following two tasks. - invalidity search: each participant group searches a collection for the patents that can invalidate the demand in an existing claim. - patent map generation: the patents associated with a specific topic are organized in a two-dimensional matrix.
A07-3	Text Summaritization Challenge 3	FUKUMOTO, Jun'ichi(Ritsumeikan University), KATO, Tsuneaki(University of Tokyo), MASUI, Fumito(Mie University)
A07-3	In QAC2, we have defined three kinds of subtasks of open-domain question answering using four year newspaper Japanese newspaper articles: five ranked answers, one list answer set for several related questions.
A07-4	Overview and evaluation results of 'NTCIR-4 WEB'	HIRAO, Tsutomu(NTT-CS), OKUMURA, Manabu(Tokyo Institute of Technology) , FUKUSHIMA, Takahiro(Otemon Gakuin University), NANBA, Hidetsugu(Hiroshima City University)
A07-4	We have evaluated multiple document summarization on important sentence extraction, redundant reduction, and readability and content of abstracts.
A07-5	Presentations by Selected NTCIR Task Participants	EGUCHI, Koji , OYAMA, Keizo , AIZAWA, Akiko , ISHIKAWA, Haruko
A07-5	'NTCIR-4 WEB' mainly aims at comparing effectiveness of information access systems for large-scale heterogeneous information on the Web, and constructing Web test collections. We report on several kinds of evaluation models, caputuring actual use of the Web from various aspects.
A07-6	Presentations by Selected NTCIR Task Participants
A08	Net-Commons Project	ARAI, Noriko , NTTDataPocket Co.Ltd
A08	Arai and NTTDataPocket Co.Ltd constructed an information portal platform, named Net-Commons. Net-Commons is a new generation information portal, which can be used as a learning management system, a contents management system, and as a groupware. National Institute of Informatics started Net-Commons Project in 2003 to promote Net-Commons as the new generation information portal platform.	ARAI, Noriko , NTTDataPocket Co.Ltd
A09	New Expansion of Digital Silk Roads	ONO, Kinji , 山YAMAMOTO, Takeo , KAMIUCHI, Toshiro, ANDRES, Frederic , KITAMOTO, Asanobu , SATO, Sonoko
A09	NII launched the Digital Silk Roads Project in corporation with UNESCO as an international joint research project at NII in April 2001. In December 2001, the Tokyo symposium was jointly organized by NII and UNESCO. The Tokyo Declaration was issued. The second international symposium for Digital Silk Roads Project was held at Nara in December 2003. A memorandum of understanding ( MOU ) between UNESCO and NII was signed by the Director Generals of UNESCO and NII. The Digital Silk Roads Initiative Framework ( DDSRIF ) was set up under MOU to undertake a coordination mechanism for promoting the Project. Bamiyan Remains in Afghan before the blasting were archived as Bamiyan Digital Museum. Digital archiving of various historical documents in the possession of the Oriental Library ( Tokyo Bunko ) in Tokyo was carried out. An Advanced Scientific Portal for International Cooperation on the DSR ( ASPICO-DSR ) to provide the platform for collaboration of Digital Silk Roads Study was developed.

Special Exhibition

Exhibition No.	Title	Division/Author
B01	Associative Information Access brings a Vivid Imagination	TAKANO, Akihiko
	GETA provides the scalable association computation which facilitates the dynamic and associative links between various databases. Users can enjoy creative rambling in the chaotic universe of information. In this corner, we introduce three information services equipped with this associative search facilities: Book Information Navigator WebcatPlus, Cultural Heritage Online in Japan, and BookMap.
B02	From Data Input to Selling in Nihon-no-Furuhon-Ya ( Japanese Old Book Shop System )	Tokyo Antiquarian Booksellers Cooperative
	We introduce the flow in Nihon-No-Furuhon-Ya ( Japanese Old Book Shop Sytem ) were the old book shops input the bibliographic data and the customers purchase the books. We also display the appearnce of the Kosho-Kaikan ( Japan Antiquarian Booksellers Hall ) which is the base of physical distribution information of ols books.
B03	Books.or.jp: the Site You can Get the Latest Information about the Books Available Now	Japan Book Publishers Association
	http://www.books.or.jp/ "Books" database contains 650,000 bibliographic records of books which were published in Japan and available now. The database is updated daily reflecting the latest bibliographic data provided by many publishing companies.
B04	Designing New Chiyoda Public Library	TSUJI, Keita , NEGISHI, Masamitsu , YAMAMOTO, Takeo , TAKANO, Akihiko AIZAWA, Akiko , KONISHI, Kazunobu , KAWASE, Masayuki , MINAI, Hideki, UOSAKI, Katsuya
	NII was entrusted with a research on the desirable services of new Chiyoda Public Library. We proposed to (1) provide electronic resources more, (2) lay stress on reference services, (3) support lifelong learning and local residents, (4) cooperate with the Kanda Jimbocho old-book town and (5) utilize information technology ( RFID, mobile phone, electronic book reader, Web, portal site and e-mail ).