No abstract

The nehvork version of question-answer processor has been developed and tested to be applied in the field of Computer-Aided Design. The concept of "Question" is understood and defined as a mismatch between the potential experience, necessary to implement a project, and the real applied experience. Existing mismatch (question) works as an active cause for decision-making control. Special questionanswer protocols are used for work with "cause-questions". The relevant answer to the "causequestion" is a result o f question processing, which controls design process and provides it with suitable information. The nehvork version is completed according to the component approach. The sofhvare has a special mechanism reflecting the designers' group structure. It is easy to install and set, it has a clear, simple and logic interactive interface.Index TermsAolleborative work, decision-making, humansomputer interaction, question answering INTRODUC~ONHuman-computer interaction involves millions of computers and users and their number will only be increased. A lot depends on how much such interaction is coordinated with human nature, in particular its (positive and negative) influence on perception of the world. The degree of such coordination depends on how much humancomputer interaction uses mechanisms of interaction of a man with the world. Interactions of the man with the environment have intellectual character, i.e. use the natural intellectual interface. The artificial intellectual interface should be built and used so as to inherit the essence of the natural intellectual interface. The essence of the natural intellectual interface is determined by a phenomenon of consciousness, which has a question-answer character.Last years we could observe the growing interest to question-answer processes and systems (more shortly QA). Many important scientific and practical results for the QA were obtained in several domains such as: the search relevant information in distributed, multimedia, multilingual and multi-agency data sets [l]; automated learning in traditional and distance education [2]; automated design and decision-making in different areas [3]. But despite more than 40 years activities, we are "just beginning to explore question-answering as a research area" [4]. The main directions of the QA research are named in the roadmap paper [5] and focus on: question a). 0-7803-8278-11041$20.00 022004 IEEE 452 taxonomies, its understanding, ambiguities, and reformulations; context and data sources of QA; real time and interactive QA, advanced reasoning and user profiling for QA. The paper considers some new decisions in the application of QA-reasoning but coordinated with [4] and [ 5 ] .The approach to human-computer interaction is submitted below on the basis of modeling question-answer activity. Our research leads to the development of the question-answer processor (with the client-semer structure) that supports collaborative activity in human-computer environments. CONCEPTLJALBASBThe obligatoly account and constrnctive use of ...

The key problem of the successful development of a software intensive system (SIS) is adequate conceptual interactions of stakeholders at the early stages of designing. Success of development can be increased with using of Artificial Intelligence means including models of reasoning. In this paper a number of question-answer means (QA-means) for decisionmaking (DM) is suggested. The base of such means is a set of the DM tasks for each of which the QA-model with the typical architecture structure is being built and used. Suggested and implemented means are embedded to the QA-processor adjusted for automated designing the SIS. INRODUCTIONDesigning of the Software Intensive Systems too often gives the results which are not corresponding to the planned expectations. The significant number of the SIS developments either are being stopped, or are being exceeded planned time and/or finance, or reach the end in the poorer version (Charette 2005). The degree of success (expressed in percentages of the number of projects, coming to the end according to the initial plans) is extremely low (about 35 %) The named kind of works is impossible without using the strict technological discipline in the DM acts which are being usually fulfilled collaboratively in the corporate network. It is important to notice that the collaborative decision-making process is a danger source of mistakes caused by misunderstanding and other problems of using the joint intellectual activities. All of these problems decrease the degree of the successfulness of designing the SIS. Therefore special means are needed to use for rational integrating the intellectual activities of designers in the corporate net during solving the project task including the DM tasks. It is reasonable to choose useful means for integration from a set of relevant AI means and includes them to the instrument of designing the SIS. First of all the relevance must be estimated in the context of intellectual capabilities such as "consciousness" and "understanding" used for control the consciousness processes. When we choose or build the relevant AI means we need to remember that the base form of consciousness work revealing is a reasoning and the dialogue (question-answer process in the brain structure) is a nature of reasoning. One way for modeling collective (collaborative) "consciousness" is to create the question-answer model (QA-model) of collaborative reasoning which is reflecting consciousness process. But for feedback the joint QA-model in any its current state must be open for the interactive "pressure" on the brain of any member of collective "consciousness" if it is useful for solving tasks. In this case individual consciousness will be combined with the interactive "pressure" of the QAmodel of collaborative reasoning (collaborative consciousness). The QA-model of collective consciousness can be used for creating the model of collective understanding which must be useful for checking intellectual activity applied to solving tasks. This paper presents the QA-model...

There are many different tasks which are to be solved on a sea vessel on the base of the primary information from radar stations (RSs) about moving objects in the vessel surrounding. Frequently enough such information is being entered from several RSs with different characteristics. The redundancy of informational flows is used for achieving many positive effects, for example for increasing the adequacy of vessel environment models. The processing of such redundant information is named as data fusion. The article presents the multi-agent system (MAS) for solving the data fusion task on the sea vessel. The suggested MAS uses the program agents not only for dynamic objects but for units of informational flows as well.

No abstract