[To Research Bulletin of the NACSIS page]@ [To NII publication page]@ [To NIIJournal page](Japanese)

Research Bulletin of the NACSIS, No.11 (1999.3)

[last updated: 1999.6.14]
Hiroshi INOSE
Preface (Japanese)

1. Noriko KANDO
NTCIR (NACSIS Test Collection for Information Retrieval systems) Project : Based on chronological analysis on Test Collections used for Information Retrieval Laboratory Testing (Japanese)

ABSTRACT This paper reports the outline of the NTCIR (NACSIS Test Collection for Information Retrieval systems) project. Based on the chronological analysis and discussion about the data used for information retrieval laboratory testing, we discuss the fundamental policies, characteristics and implication of the project. NTCIR-1, its first test collection, contains ca.330,000 documents, more than half are Japanese-English paired. A search topic contains detailed narrative including term definition, relevance judgment criteria, the purpose of the search, and background knowledge, which are thought to be helpful for relevance assessment. The Workshop, which started from November, 1998, obtained thirty-one participating groups. The relevance assessment of the NTCIR-1 will be done through the workshop. Future studies on test collection and evaluation of information retrieval are discussed.

On the Study of Dynamics of Terminology: A Proposal of a Theoretical Framework (Japanese)

ABSTRACT The purpose of this paper is to clarify concepts referred to by 'dynamics of terminology' and to outline a possible theoretical framework within which the dynamics of terminology can be described in such a way that the description can be interpreted as a study of terminology in itself de jure. Many efforts have been made so far, in vain, to make the study of terminology appear an independent discipline, by examining the 'theoretical' or 'essential' nature of terms as an empirical object. This kind of enterprise is, of course, logically doomed to fail, as the status of the theory of terminology is related to the 'theory' of terminology and not to terms as an empirical object. As a preparation for the purpose of the paper, therefore, I first clarify the distinction between quid iuris and quid facti concerning the study of terminology. On the basis of this distinction, I then affirm the status of the study of dynamics of terminology in its own right, and then sketch a possible means of describing the dynamics of terminology.

3. Masaharu YOSHIOKA
Research on Keyword Role Analysis based on the information of title and abstract (Japanese)

ABSTRACT There are several types of research articles; e.g., survey articles, those that propose new technologies, those that apply existing technologies to unsolved problems. In this paper, I propose a methodology to categorize these types of research articles by analyzing roles of the author added keywords. First, I categorize verbs that are frequently used in research articles into five role categories. Second, I propose methodology to analyze roles of the keywords by using the categories of verbs. To validate this methodology, I apply it to the real research articles and explain its result.

4. Eisuke NAITO
One Aspect of Information Policy of Japan and Standardisation Tasks (Japanese)

ABSTRACT Based on a framework proposed by N. Moore, aspects of Japanese information policies are reviewed. Development of Japanese information policies is introduced as before G7 (Group 7), after G7, and the Program for Advanced Information Infrastructure (1994-). Pattern, a dirigiste model, is seen between government promotion and reaction of the private sector. Mediacy promotion by the Japanese Ministry of Education is shown as an example of government action towards primary and secondary education. A private sector activity (AFSIT, MLIT) is introduced on achieving multi-lingual equality, quite different from that of European needs and much in demand in Japan as well as in other Asian countries.

Grammatical Categories for Constituent Elements of Japanese Nominal Compounds with Special Reference to Technical Terms (Japanese)

ABSTRACT Grammatical information for constituent elements of nominal compounds are needed for analyzing nominal compounds, especially those of technical fields. In this research we analyzed the grammatical categories suitable for nominal compound analysis, as it is difficult to apply conventional grammatical categories defined with respect to words in sentences to the analysis of nominal compounds.

6. Kotaro NAWA
Scholarly Communication and Copyright Law in the Electronic Environment (Japanese)

ABSTRACT Researchers have a traditional idea that they should hold scholarly documents in common and they should use the documents freely, in spite of restrictions by copyrigt law.
Electronic publishing will change the public documents into proprietary works and will cast off the traditional idea into the market.

7. Teruyoshi HISHIKI, Jun-ichi TSUJII, Kazuhiko OHE
Extraction of Medical Knowledge from Japanese Text (English)

ABSTRACT With the increase in knowledge-based medical computer programs, automatic extraction of medical knowledge from natural language sources would be necessary. We studied the Japanese medical resources to find out the literal and syntactic patterns where the relationship between symptoms and their relevant body parts is represented. We initially assumed that the patterns can be classified in three ways. We evaluated our hypothesis by examining the sentences in the sections for clinical manifestations of infectious diseases from Japanese electronic medical textbook. Our experiments have shown 66% recall and 92% precision by using these three patterns only. Though our method is based on a simple technique, we think that our method would be useful to acquire various knowledge from Japanese medical textbooks in the future.

8. Vuthichai AMPORNARAMVETH, Eiji IKOMA, Akiko AIZAWA, Kinji ONO
Interactive Large-Scale Pictorial Information Retrieval System for the Internet (English)

ABSTRACT This paper describes development of a network-based client/server system designed to support information retrieval of high-resolution image interactively. Through an interface implemented using a JAVA applet, the system allows any Internet users with a JAVA-support browser to navigate high-resolution images in real time. User satisfaction is improved by transferring image progressively and use of image cache on both server and client sides. The system can also be customized to provide supplemental image-specific data, image configuration window, balloon help, and image map features. These features make it possible to create wide-range of large-scale image applications on top of the system.

9. Katsumi MARUYAMA
A simple distributed active-object system for Java Virtual Machine (Japanese)

ABSTRACT Existing distributed processing systems, such as distributed database systems and file systems,are mostly based on the Client-Server model. However, in future systems,the peer-to-peer distributed processing model will become more important,where each object works autonomously and concurrently,communicating each other in distributed environment. An active object,which has its own thread to execute its methods and can communicate each other asynchronously over network, is very suitable to implement the latter. This paper explains an active object distributed processing system using Java.

10. Haruki UENO
Knowledge Modeling Based on Cognitive Science for Autonomous Human-Type Service Robot (Japanese)

ABSTRACT This paper describes the background concepts and key methodologies of the on-going robotic project, i.e., the autonomous human-type service robot HARIS, especially knowledge-based information modeling based on Cognitive Science. The HARIS robot consists of two parts, which are the HARIS arm and the HARIS system. The HARIS arm is a robot arm itself which consists of 5-finger hand with 17 d.o.f. and 3-joint arm with 8 d.o.f., and has local intelligence for manipulations with 178 tactile sensors. The HARIS system, a software system which realizes high-level intelligence, consists of a 3D vision, an intelligent task scheduler, an intelligent motion scheduler/controller and two knowledge bases, i.e., the world model and the task model. The world model maintains the knowledge to understand an environment by model-based vision and create the mental image of it. The task model maintains the knowledge on the primitive tasks to generate a sequence of tasks to achieve a request at verbal level. The important characteristic of the HARIS system is that the vision and task scheduling method is based on Cognitive Psychology, i.e., a human's way of vision and scheduling is applied. The current system receives a request such as "Put a red cup onto a white tray", understands an environment, generates a task sequence at verbal level such as "reach", "grasp", "lift", etc., and generates a sequence of commands to control the robot with simple monitoring. The long range goal is to provide a paradigm for human-robot symbiosis. The frame-based software platform ZER++ plays an important role in developing the HARIS system.

11. Hironobu GOTODA
3D Scene Reconstruction from Multiple Views Assisted by Rendering Hardware (Japanese)

ABSTRACT Image-based rendering is a method to integrate multiple views of a 3D scene for computing and generating the view at an arbitrary eye position and view direction. Recently, the method has received much attention, as it provides one of the easiest way of realizing 3D navigation systems, such as virtual reality systems. However, since the method cannot utilize the depth information of the scene, the generated views are often skewed, especially when the distances between the objects and eye positions change.
To address this problem, we propose a new method of 3D scene reconstruction, which is featured to recover the depth information from multiple views. The method takes advantage of the voxel-based 3D space representation to perform view matching in parallel. Although there exist many other methods, such as stereo matching, for reconstructing 3D information from 2D images, they often suffer from the computational complexity, i.e., real-time reconstruction is almost impossible with such methods. On the contrary, the proposed method can be accelerated by the assistance of 3D rendering hardware, and will be more suitable for wider range of applications.

12. Jihad BOULOS, Kinji ONO
Cost Estimation of User-Defined Methods in Advanced Database Systems (English)

ABSTRACT In this paper we present a novel technique for cost estimation of user-defined methods in advanced database systems. This technique is based on multi-dimensional histograms. We explain how the system collects statistics on the method that a database user defines and adds to the system. From these statistics a multi-dimensional histogram is built. Afterwards, this histogram can be used for estimating the cost of the target method whenever this last is referenced in a query. This cost estimation is needed by the optimizer of the database system since this last needs to know the cost of a method in order to place it at its optimal position in the Query Execution Plan (QEP). We explain here how our technique works and we provide an example to better verify its functionality.

13. Yusheng JI
A New Scheduling Scheme with Better Performance for Shaped Traffic (Japanese)

ABSTRACT Scheduling mechanism in switching nodes is necessary to isolate traffic while obtaining multiplexing gain in integrated services networks. The interference from other traffic can degrade the quality of service, such as mean and maximum delay, loss probability, etc. On the other hand, allowing resource to be shared fairly between traffic is another requirement to scheduler. In this paper, we introduce a new fairness criterion, which define fairness as when resource is fairly shared among traffic during a certain time period, instead trying to divide resource fairly at any instant. Schedulers based on this kind of fairness criterion can be expected to give more chance to shaped traffic, i.e., shaped traffic will have better quality when serviced under scheduler conforming to the new fairness criterion. Analytical results show that the worst-case performance of new scheduling algorithm CBFQ is not bad than traditional WFQ, also numerical results show that shaped traffic can be expected to have better statistical performance.

14. Weiping ZHAO
Implementation and Performance Evaluation of Object-Oriented Hotspot in AutoCAD Environments (English)

ABSTRACT An AutoCAD database is an object-oriented database (OODB) that can be used to constitute an information retrieval system such as geographical information system (GIS) and mapping system. Although the AutoCAD database is designed reasonably and rationally, it has some insufficiencies. For example, it takes time to search entities from a large database because all entities are equally owned by a table which means that to search a specific entity must retrieve the whole table space. Meanwhile, AutoCAD Runtime Extension (ObjectARX) provided with AutoCAD makes it possible to improve the insufficiencies of AutoCAD database by developing new objects and new search methods.
In this paper, a custom object called hotspot entity whose graphical representation size can set independent from the zoom depth is introduced to provide a tools to place anchors or markers in a drawing, and a fast search method using a dictionary containing the hotspot's search information is proposed to retrieve a specific hotspot entity from an AutoCAD database. A hotspot toolkit consisting of the proposed hotspot entity and fast search method is developed to design a more efficient and more sophisticated information retrieval system based on the AutoCAD OODB.

15. Shunji ABE
An Estimate Method of ATM Cell Quality based on Traffic measurements - Basic idea - (Japanese)

ABSTRACT Putting ATM network to practical use is very rapidly progressing. In this situation, it is very important to design ATM network parameters such as bandwidth and topology efficiently and economically, and also an efficient and an economical operation of an ATM network is important. On this account,it becomes it momentously to know quality of an ATM network in use precisely.This paper describes a cell quality estimate method of an ATM network inoperation. This method is as follows. The first, a multiplexed cell traffic stream is approximated by a model that generates cells of two kinds speed alternately from traffic measurements of an ATM network in operation. The next, cell loss rate is estimated by the approximated model. In order to realize this method, it is necessary to determine four probability distributions(each of cell interarrival distributions for two kinds speed and their duration distributions) for the approximated model. As the first step toward realization of this method, this paper considers what kind of distribution is propriety. we show that when each of duration distributions is 2-stage hyperexponential, if 2-stage hyperexponential and exponential are selected as cell interarrival distributions respectively, cell loss ratecan be estimated by practical accuracy.

16. Takayuki FUJINO
Implementation of Administrative Routing Information Managing System for Keeping Routing Stability (Japanese)

ABSTRACT Recently, network resources, such as bandwidth, scale, have been improving since the number of Internet users increase. As a result, network scale becomes larger with more routers, and network topology become more complex. This means factors should be considered by routing operators are also complicated. Furthermore, wide-area networks are often managed by several operators, so the accident such as elimination, modification of routing configurations might be happend.
It is required that administrative routing informations are managed and shared by operators in order to keeping routing stability. This paper discusses which informations should be managed and introduce the implementation of administrative routing information managing system.

17. Frederic ANDRES, Marielle JELADE, Kinji ONO
New Trends in Distributed Information Engine (English)

ABSTRACT Hypermedia systems are one key class of applications of modern information system. They represent a new infrastructure and environment that has been created by the integration and the use of computing, communications, and hypermedia document and media data management on a global scale. Both, the explosion in the volume and the complexity of digital information impose new features from these systems. Hypermedia systems have to tackle several important issues such as uniform data storage, interoperability, customizability and active behavior. These systems manage hypermedia documents for different application domains such electronic library, environment control or workflow system. Hypermedia documents refer to inter linked multimedia objects which can be structured or unstructured text, images, video and audio information.
In this paper, we survey the new trends in distributed multimedia information systems in keeping with our system AHYDS, which enables to store and to retrieve an extensible set of hypermedia documents according to the requirements of the applications. Mainly, research works have addressed three evolving domains: data processing technology, the operating system capabilities (e.g. supervision, load balancing and fault tolerance), and the agent technology.

18. Masaki NISHIZAWA, Sumio KAKINUMA, Yuan SUN, Masaharu YANO
A Comparative Study on Information Science Research Activities in Japan and the United States (Japanese)

ABSTRACT In order to maintain the superiority of Japan in information technology and information science areas, it is important to explore the reason why the United States leads all the countries in the world in many of these fields. In this paper, we give some statistical comparisons between Japan and the United States based on the number of researchers, the amount of research funds, and the number of scientific publications. The results obtained are as follows: (1) In information science fields, Japan has relatively the same number of graduate students as the United States, but less than half the number of researchers of the United States in normalizing for the population size of the two countries. (2) The University research funds by Japanese government is about one-fifth of those by the U.S. federal government as a share of GDP. (3) By the analyses of numbers of scientific publications derived from Science Citation Index (SCI), it is revealed that Japan has a high priority in "Information Systems", "Hardware and Architecture", and "Theory and Method", with relative little effort in "Software, Graphics and Programming" area, on the other hand.

19. Sumio KAKINUMA, Kenneth PECHTER
University-Industry Research Collaboration in Japan: Industry Dependence on Academic Research (Japanese)

ABSTRACT How closely do academia and industry in Japan collaborate in the research process? We address this question through the window of coauthored papers. According to our analysis, there is no question that the coauthoring of scientific and technical papers between Japanese industry and academia in Japan and abroad is prevalent and has been rising over the 1981-96 period covered in this study. In 1981, 23% of all industry papers were coauthored with academia. By 1996, this had become 46%. This rise has been enough to overtake the percentage of industry articles authored within the firm, which fell from 70% to 43% over this period. Our analysis provides evidence that there is indeed a strong and increasing research interaction between industry and academia in Japan.

20. Toshihiko NOZUE, Mika KOSHIZUKA
'Fee or Free' Problem and User Education for New Information Services in the Public Libraries: A Case Study on Some Public Libraries in U.S. and Canada (Japanese)

ABSTRACT New information services such as CD-ROMs, online databases and internet are becoming popular in the public libraries. Fee or free' problem and user support (user education) are two of many problems which have to be considered concerning to such services. In this article, a case study on four libraries in U.S. and Canada is reported.

21. Akitoshi EDAGAWA
A Study on Enhancement of the Service of the Information System among Art-Museums (Japanese)

ABSTRACT The cultural facilities, especially, art museums were established independently and managed. As for their exchange, the loan debt of the display article has provided the contents of the exhibition for the other museum. However, recently, it is possible to divide the building of an information system among the art museums, which used an Internet-network mainly in the advanced countries in the line. The active contents at the art museum became widely exhibited to the use person, too, using the Internet. With the development of the information communication technique, the virtual-art-museum which doesn't have a building in future, too, becomes able to be thought of.

22. Fumio KODAMA, Masaharu YANO
The Creativity of Teams and the Management (Japanese)

ABSTRACT We measured the creativity of R&D teams in a firm by the subjective judgements of peers. Mainly statistical analysis shows that the creativity as a team character exists, and that the variety or the heterogeneity of the team member generates the creativity of the team. In other words the team creativity is made by management. The individual analyses of the laboratory, the department of systems engineering, and the department of information processing show that the mission or the substance of the department determines the relation between the variety and the creativity.