Human Information Interaction Research Articles

Human-Information Interaction with Knowledge Associates

Under the Novel Intelligence from Massive Data program, Battelle is working in partnership with Stanford University's Knowledge Systems Laboratory and IBM's T.J. Watson. Research Center to develop a suite of technologies for knowledge discovery, knowledge extraction, knowledge representation, automated reasoning, and human information interaction. We envision an integrated analytic environment composed of a collection of analyst associates called “Knowledge Associates for Novel Intelligence (KANI).” One of Battelle's contributions to this environment is the design of an Information Interaction Associate, through which analysts will interact with other knowledge associates in KANI and actively construct and manipulate explicit models of the situations, events, actions, and scenarios that they analyze and hypothesize about during an analysis task. This paper reports on efforts to date towards the design of this component and efforts to develop and encode of a model of the intelligence analysis process.

Research on information interaction and intelligent information provision mechanisms

Information systems comprise a user, who recognizes that his/her problem is inadequate for attaining some goal, a knowledge resource, which may offer useful information, and some intermediary access mechanism to the knowledge re source. The last two components constitute the information provision mechanism (IPM). The information system we are particularly concerned with is that of document retrieval where the knowledge resource is typically a set of online biblio graphic databases and the intermediary is, at present, a human search analyst. For a variety of reasons—social, technological and eco nomic—, it has become important to consider what aspects of the IPM can be performed, in an intelligent way, by computer programs. This paper reviews the methods and results of a six-year ongoing series of research projects concerned with this issue, and especially the question of how to design an intelli gent automated system interface. The specifications for this interface have been derived from a study of real life human user-human intermediary information interactions. A variety of methods has been employed, particularly functional dis course analysis, and results are: the specification for a distrib uted expert system architecture; identification and specifica tion of the necessary functions; and proposals for a problem- structure driven human-computer dialogue.