Past Projects


The following projects have been completed


1.    JST CREST Digital City Project (2000-2004)


People are starting to use the Internet, not only in business but in their daily life as well. The Internet for business needs to be a homogeneous and logical information space. On the other hand, the Internet for everyday life should be a space that is heterogeneous and contains rich information reflecting cultural characteristics of a local area. For example, patients suffering from high blood pressure do not require the Internet as a worldwide network. Instead, they need the Internet as a space that encompasses everyday life information of the surrounding vicinities.


Digital City Kyoto started in October 1998 as a collaborative research project between Kyoto University and NTT; it has been lead by the NTT Open Lab at Keihanna. The project is an experiment focusing on the future of social information space for academic and business uses. The Digital City Kyoto Prototype Forum started in July in 1999 to make studying, developing, and experimenting on digital cities easy for more organizations, not limited to NTT. The project is based on the idea that future internet societies will based in large information spaces designed for everyday life. The Digital City Prototype Forum is an organization set up to study Kyoto, and to study and to develop technical solutions for future digital cities. This forum includes universities and companies interested in developing digital city infrastructure as well as providing technology and administration to manage digital cities. Additionally, member organizations provide researchers to analyze the experiment as a whole and investigate the future of digital cities.


In July 2000, Universal Design of Digital City, a five-year project was established (project director: Toru Ishida), under the Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Corporation (JST). The objective of this project is to create digital cities as an infrastructure used and participated by all people, including the handicapped and the aged. However, the high cost of developing systems specifically for the handicapped and the aged, in return, burdens them. Therefore, it is necessary to design the Digital City from the beginning so that everyone can use this information space without any barriers. The universal design of digital cities -useable for all people- is indispensable for creating the information space for daily life. In this project, we will develop basic technologies for the universal design, subject to sending information, receiving information, and participation.

This project is promoted by JST Digital City Research Center located in center Kyoto, collaborating primarily with the Department of Social Informaticsat Kyoto University, the Faculty of Systems Engineering at Wakayama University, NTT Communication Science Laboratories, Advanced Software Technology & Mechatronics Research Institute of Kyoto (ASTEM), and Advanved Telecommunications Research International (ATRI). The domestic cooperation also includes the Department of Architecture at Kyoto University and the Digital City Kyoto Experiment Forum. The international cooperation is underway with Stanford University, California University, Paris 6 University and Shanghai JiaoTong University.


Digital Cities have used technologies such as geological information systems, virtual space and mobile computing. In our project, we carry out basic researches into the information infrastructure called perceptional information infrastructure (PII) and the software called social agents. In PII, we create an omni-directional vision network. It models the function of environmental static and human active structures to accumulate information actively. We also construct a new technology that is able to record and to search a large quantity of time series data. In the research of social agents, we design a communication environment for agents using a 3D virtual space. Within a virtual meeting space, we examine the effects of the agent's utterance and behaviors in the society, and evaluate the agent's social roles. The research of media adaptation bridges the gap between information utilized by the social agent and the information accumulated through PII. It examines the relations between various media expressions in a city and human's interpretations. Then, the technology transfers the accumulated information in accordance with the user's knowledge and sensitivity. The result of these fundamental researches is applied to security, learning about the natural environment by experience, crisis management, and so on. The field experiments are developed on the stage of real cities, which includes Kyoto.

To connect residents and visitors, one trial for developing social interaction in digital cities uses avatars in the 3D space. Avatars are controlled by humans walking in the 3D virtual space called FreeWalk. This technology allows a number of avatars to walk around the digital city in real-time. By making links between the avatars and the people walking in the corresponding physical city, we can realize communication between digital tourists and people walking in the city. As walking motion can be generated by the user's machine via a WEB browser plug-in, only the walking position/velocity and direction need be downloaded. Thus, a large number of avatars can be created rapidly in real-time. Aside from the known avatars, adding a virtual population will activate the digital city and make it more attractive. In addition, a scenario description language called Q that can control the behavior of thousands of social agents is currently under development. Agents autonomously behave under the social constraints given by the scenarios, and dynamically respond to other avatars (humans) and agents (software).


We are trying to create a group of social agents that support the various activities of participants in a digital city. We expect social agents to show their ability as coordinators for building and maintaining online communities. The playgrounds of social agents include virtual and mobile environments. In a virtual environment, social agents interact with avatars (humans) visiting the same world. In a mobile environment, social agents appear on PDAs or wearable computers to interact with people living in the physical world. Social agents can enter into communities in both environments. The service industry is a good example. Social agents can troll through vast information repositories of products and services, communicate with customers, and work twenty-four hours a day.


Evacuating people safely in the case of disaster is another example of social agents playing an important role both in digital and physical cities. In evacuation simulation, the possible roles of social agents include pedestrians, station employees, salesclerks and so on. Realistic evacuation simulations can be realized by having pedestrian agents act as people running around to escape. Such simulations would help a crisis management center to accumulate experiences and to make correct decisions, since social agents would exhibit the mistakes typical of people. In real evacuation, social agents would appear on the user PDA to provide appropriate instructions. For success of simulation, social agents have to be trusted by people. Functional accuracy to make correct decisions is not enough to be trustworthy agents. Technologies to create trustworthy agents are to be explored.


To understand the nature of social agents, we are performing a series of social psychological experiments. Effects of introducing social agents in human communities have not been well investigated. We are currently conducting experiments to see how social agents support human communication and influence human relations. Social agents can act as go-betweens among people who have different social identities such as inhabitants and visitors, young and aged, and so on. In our experiment on cross-cultural communication between Japanese and American students, the agents influenced not only the impressions of agents but also the impressions of conversation partners and the stereotypes of nationalities. For example, if the agent encouraged students to discuss political problems, the Japanese students became as talkative as the Americans. Another experiment to observe the influence of agents on human-human relations shows that social agents have the ability to affect human relations. Results of this experiment demonstrate the ability and the limitations of agents to influence human relations. In future, we may need an ethics law governing the development of agents.


2.    Intercultural Collaboration Experiments (ICE2002-2005)




The diffusion of the Internet is promoting new international alliances around the world. Responding to the alliances rapidly being formed in North America and EU, Asian countries need to build true partnerships in this area. As the first step, we plan to conduct an experiment on open source software development by multinational teams in Asia. Because one of the unique characteristics of this experimental project is the pursuit of collaboration among heterogeneous groups across country borders, in this experiment, we won't use English as a standard language. We feel that the language barrier among participants will be successfully overcome through the use of machine translation tools. Furthermore, participants of this experiment should overcome translation mistakes through intense and frequent communications with others.


The definite goal of this experiment is the pursuit of globalization with cultural heterogeneity (unlike the current trend of globalization through standards) for the next generation. Given that the first phase of globalization is standardization of platforms, this experiment targets the second phase of globalization, which respects different cultures and seeks harmony among them through intercultural (multilingual) communications.


As part of Department of Social Informatics of Kyoto University, and with the support of Japan Science and Technology Corporation, we are currently preparing multilingual communication tools called TransBBS and TransWEB for this particular research experiment, for Chinese, Korean, Japanese, Malay and English, based on available translation services. NTT MSC Sdn Bhd, NTT Cyber Space Laboratories, and Kodensha Co.,Ltd. has offered their translation services to this experiment.


In this experiment, the source programs of our multilingual communication tools TransBBS and TransWEB will be provided by Kyoto University. Based on these source codes, a team consisting of students from China, Korea, Japan and Malaysia is developing intercultural collaboration software. The first track involves Shanghai Jiaotong University, Seoul National University, Handong University, University of Malaysia, and Kyoto University. In this experiment, team members living in different countries will never see with each other in the real world, but will complete the software over the Net. Basically, they communicate with each other in their native tongue using translation services. If the machine translation fails at some point, a second language (such as English) can be used. The results of this experiment will be published in international conferences and journals.


The experiment is composed of the following two phases. In the software design phase, team members design intercultural collaboration software. For example, functions will be added as needed to overcome the problems of the prepared multilingual communication tool TransBBS and the collaboration tool TransWEB. The goal of this phase is to submit a system design proposal and prepare for the software implementation. In the software implementation and integration phase, team members implement software based on the design proposal. The goal of this phase is to complete and release an intercultural collaboration tool. Here, team members have to integrate software modules into single interface software. To achieve this goal, accurate information exchanges among members are required.


The communication space in this experiment consists of our multilingual communication tools TransBBS and TransWEB. Team members will their functions to yield effective intercultural collaboration software. Basically, team members communicate with each other in their native tongue via TransBBS and TransWEB. If, however, the machine translations hinder the discussion, a second language (such as English) can be used in combination with their native tongue.


This experiment is sponsored by Department of Social Informatics, Kyoto University, Japan Science and Technology Corporation (JST), NTT Communication Science Laboratories, and supported by Information Processing Society of Japan (IPSJ), Korean Intelligent Information Systems Society (KIISS), IEEE Malaysia Section, Shanghai Computer Society (SCS).


3.    Human Centered Semantic Web (2003-2005)


The Semantic Web attempts to enable software agents to process information and services provided on the Web. Various researches to describe the semantics of information and services are now in progress. Current fundamental Semantic Web technology, however, focuses on computation-centered ones based on description logic, theorem proving, situation calculus, planning and so on. Since these technologies do not pay attention to humans who actually create and use Web contents, the high-quality, widespread use of this technology may not be guaranteed. In order to bring real benefit for humans out of the Semantic Web, we need to develop human-centered technologies that will bridge the gap between the human use and the computation-centered mechanisms.


This research will attempt to fill the gap that exists in the following three components of the Web. A) Interface: We will fill the gap between the direct manipulation interface of the current Web and the agent interface of the Semantic Web. The research includes automatic acquisition of Web Service scenarios for agents using the example sequences of human operations. B) Contents: We will close the gap between the current Web contents written in HTML and the formal Semantic Web description. The research includes automatic extraction of ontology from a large volume of HTML texts and automatic insertion of semantic tags into HTML content. C) Process: We will fill the gap between the actual process of human Web usage and the formal process model postulated by the Semantic Web. The research includes the examination of the human-in-the-loop process for generating composite Web services from existing services. To do this, we will conduct real scale experiments such as Digital Cities, Intercultural Collaboration Experiments, and E-commerce. Through these experiments, we will extract process ontology such as collaboration and negotiation ontology.


In the August 2002 edition of IEEE Spectrum, the Semantic Web was chosen as one of the five blockbuster technologies to appear in the next five years. But voices that predict the failure of the Semantic Web are equally strong. When we look at the future of the Semantic Web, we cannot help but notice the distance between the ideal and the reality. That is, there is a gap between the goal that Tim Berners-Lee of W3C proposed in the article of Scientific American, where the Semantic Web is described as not an AI technology, and the actual researches represented by DAML, OWL, DAML-S, where computational AI research is conducted.


Considering that humans use today's Web only through direct-manipulation interface, seamlessly providing both the computation-centered mechanism and human-centered interface is necessary to guarantee the high-quality and widespread use of the Semantic Web. Researches with such human-centered perspective are not yet prevalent, but will become active in a global scale in the future. This research will play a leading role in the human-centered research area.


In the United States, DARPA and NSF are making large investments in the Semantic Web research, while IST (Information Society Technologies Programme) is taking that role in Europe. Since Western countries are now heading the Semantic Web research, there is little room for Asian countries to make contributions by taking a similar approach. Human-centered research, however, has yet begun, and there is a great chance that Asian research communities' practical and empirical approach will benefit the world.