Informatics Institute Annual Report 004 Chapter introduction

Human Computer Interaction for the Blind

In order to gain insight in the barriers towards optimal user experience for blind and visually impaired our research aims to develop an international online community for blind and seeing children. This research focus offers us the opportunity to push theory and technological innovation further to cater for solutions to the problems we encounter. These innovations are thought to offer contributions for the field of Human Computer Interaction in general and contribute towards bridging human sensory capabilities with technology. In 2004, our efforts have focused on requirements analyses of blind and seeing children to identify differences in requirements and user experience. We find that blind children build different cognitive models concerning Internet and computer technology than seeing children do and these differences influence interaction.
Cultural aspects of E-government

In previous studies investigating the effects of culture on design preferences, attitude towards computers, understanding of interfaces, use of e-government systems, we have measured cultural value dimensions and have tried to relate them to design preferences, user attitudes, user experience or usage behaviour. In a recent study we investigated multi-cultural user experience of e-government applications for SME’s, In another study we tested attitude towards computers in the Akshaya project (a pilot project in rural India to make at least one person in every household computer literate). In these studies, cultural dimensions were not found to correlate with aspects of attitude or user experience. Our findings suggest that socio-economic factors such as knowledge of language, education and computer experience are better factors to predict user preferences and attitudes. Even though these factors offer the ability to model a particular target user group, it still leaves the question about the way in which cognitive models of the real world are shared within a culture and whether modeling of users on a group or cultural level is possible. Not only to assess international and cultural aspects of Human Computer Interaction but also to inform theories on collaborative systems. In investigating cross-cultural user behaviour we will need to build a complete and more complex model of the user and the context in which technology is used.

E-government for multicultural entrepreneurs.

The city of Amsterdam is developing many new ways to improve service delivery to citizens and entrepreneurs in the city. Recently, a virtual business counter, het ondernemersloket, has opened in Amsterdam. We investigated how users with different ethnic and cultural backgrounds experience the use of this virtual counter. The main findings indicate that language skills and virtualisation of the company of the user affect the use of the virtual counter more than cultural background and other demographic variables.

ICT for development

Other research focuses on poverty reduction in developing nations and the role of technology in this process. Previously, a project was briefly mentioned where technology acceptance was evaluated for people in rural India who are taking part of a computer literacy project. The research offered insights in the impact of technology on economically disadvantaged communities. Another study investigates access to education for children in developing countries. This research is in collaboration with the NIIT institute of India (KIOSK project) and involves the development and placement of publicly accessible computer kiosks in rural India. Our research investigates what knowledge children develop by using computers without any assistance. The research provides insights into the knowledge needed to make optimal use of computers and the internet and what schooling and knowledge could potentially be provided through the kiosks.

Learning in simulation environments

This project investigates factors that influence learning processes in simulation environments such as KM-Quest. The main results of the research are:

1. Contrary to the literature (cf. De Jong & Van Joolingen) the gaming –simulation environment KM Quest for the domain of Knowledge Management does promote learning success. Students acquire declarative and procedural knowledge as a result of playing the game.

2. The prospective measurement of meta-cognition (in which meta-cognition is regarded as a more or less stable trait or disposition) by means of a self-report inventory, does not influence the acquisition of knowledge and skills

3. The retrospective measurement of meta-cognition (in which meta-cognition is viewed as a transitory state) by means of a self-report inventory, does influence the acquisition of knowledge and skills. Students that report to be weaker in using meta-cognitive skills, achieve more learning success than students that report to be stronger in using meta-cognitive skills. The assumption is that KM Quest, that is strongly rooted in constructivist principles, is especially beneficial for students that are weaker in monitoring and controlling their learning behaviour.
4. The addition of a task model to the learning environment that prescribes how to solve KM problems is of no influence on learning achievements or use of meta-cognitive skills.
2005 and beyond
In 2005 further research in the ICIS-CHIM project will investigate the interaction necessary for user involvement in automated filtering or information distribution. In complex collaborative dynamic environments such as crisis control situations, information overload is countered through adaptive information filtering. These filtering systems are based on machine learning technology and mix automatic and manual profile generation. The way in which users in complex dynamic settings can collaborate and collectively train the information distribution system will be one of the focus areas of research. Affective aspects of interaction will be investigated as effectiveness and efficiency of the system will depend on user trust and acceptance.
A new research project proposed involves user interaction and interface preferences for multimodal Virtual Reality visualization. Virtual environments support a variety of modalities ranging from non-immersive desktop representations on a conventional PC or a PDA, to fully immersive CAVE-like environments and augmented reality (AR). The range of functionalities supported is variable and can focus on projection and analysis of complex data (MRI scan based visualisations of human physical conditions) or virtual meeting places where remote social interaction is realised through virtual presence in a virtual space. The selection of an appropriate modality in accordance with the user’s tasks, preferences and personal features is increasingly important. The research will investigate to what extent task-orientation, user culture and socio-psychological factors influence user preferences for interaction in different projection modalities.
Ongoing research concerning special user groups will lead to developments of prototype systems to facilitate interaction between blind and seeing users. E-learning tools to facilitate unsupervised learning in disadvantaged communities and the development of a model of culture as well as a model of emotion to effectively investigate the effects of culture and the effects of emotion on human computer interaction. Data collection tools will be developed from these models and administered in a large scale international empirical study.
Qualitative Reasoning
Keywords
Knowledge capture and communication, Knowledge simulation, Conceptual knowledge, Common-sense reasoning, Question generation, Diagrammatic reasoning, External representations, Concept Maps, Explanation.
Main theme
Qualitative reasoning is an area of Artificial Intelligence that is concerned with the construction of knowledge models that capture insights domain experts have of systems’ structure and their behaviour (functioning). Qualitative reasoning provides a rich vocabulary for describing objects, situations, relations, causality, assumptions, and mechanisms of change. Using this vocabulary it is possible to capture conceptual knowledge about systems and their behaviour and use that knowledge to automatically derive relevant conclusions without requiring numerical data. Qualitative modelling uses a compositional approach to enable reusability. This is achieved by constructing libraries of partial behaviour descriptions (model-fragments) that apply to the smallest entities relevant within a domain. As larger systems are built from these basic elements, reasoning about the behaviour of larger systems means combining the behaviour of these elements. This prevents having to develop dedicated models for each system encountered. Our main research questions include:

• 
  Constructing knowledge libraries for new domains
  
• 
  Distributed and collaborative model-building
  
• 
  Automated question generation
  
• 
  Explanation generation and automated model debugging
  
• 
  The application of QR technology to education
  
• 
  Diagrammatic languages for knowledge visualization
  

2004 Results