Ebook Supporting Group Work in Crisis Management: Visually Mediated Human-GIS-Human Dialogue
Crisis management is a time-critical and collaborative activity that involves multiple individuals and organizations sharing information, expertise, and resources in support of rapid situation assessment and decision-making activities. Geospatial information is fundamental to supporting crisis management at scales from the local to the global. The potential roles of Geographical Information Systems (GIS) in providing this information to crisis managers have been detailed repeatedly (Cutter, 2003; Mondschein, 1994; Kumar et al. 1999).
In spite of the potential, several impediments exist to widespread utilization of GIS in crisis management activities (Zerger and Smith 2003). These include limited data and software interoperability, lack of mechanisms for immediate integration of real time geospatial data into the crisis management process, difficult to use GIS interfaces that force crisis managers to rely on GIS technicians to provide necessary information, and lack of support for same and different-place group work.
Our research addresses the latter two issues. Specifically, we are developing a conceptual approach to natural, multimodal, dialogue-enabled interfaces for GIS that will make GIS more responsive to nonspecialist user needs and to supporting group work with geographical information. In this paper, we focus on map-mediated, dialogue-enabled human-human communications in the same place and at the same time.
Central to the dialogue-enabled interaction is a dialogue management component which serves as a semantic mediator and information flow controller between human users and the GIS. This dialogue component (initially developed to support human-system interaction) has been extended to provide mechanisms to support group work. We have implemented aspects of the approach and have begun to assess its use and usability(MacEachren, Cai et al. 2003; MacEachren and Brewer 2004; Cai, Wang et al. 2005; MacEachren, Cai et al. 2005; Cai, Sharma et al. forthcoming).
In the previous work cited above, we have focused on interpretation of multimodal requests from users to a GIS. The goals in this work have been to: model the strategies humans use to specify their geospatial information needs through language and gesture and using these models, to create a natural interface that supports easy information retrieval from a GIS. Our more general goal is to support more efficient and richer dialogue between users and the system and among users enabled by the system. Through the work carried out thus far, we have recognized a need to develop additional strategies that leverage the flexible map-based display provided by GIS.
While there has been considerable past attention in multiple domains to both dialogue-based interfaces and group work with GIS, neither body of research had directed much attention to the importance of visual display, or to maps specifically, as a component of dialogue or a mediator of group work with GIS and there has been limited work directed toward integrating advances in dialogue-based interfaces with that in group-enabled GIS. We believe that treating the visual, map-based display as part of an ongoing dialogue (rather than a simple repository of information) may be the key to computer-supported group work with geospatial information. In this paper, therefore, we focus on the use of visual representation as a mediator for dialogue, both between a human and the system and among humans.
In the next section, we present a theoretical framework for conceptualizing the roles of visual display in facilitating effective human-human collaboration involving geographical information. The framework is based on cognitive and social theories relevant to human-machine and human-human communication, with an emphasis on distributed cognition and the role of shared artifacts in facilitating group work. Then, in section 3, we describe our prototype system, DAVE_G (Dialogue-Assisted Virtual Environment for Geoinformation), and articulate the associated cognitive-semiotic principles towards enabling humanhuman task-oriented dialogues.
In section 4, we then outline the computational framework for our dialogue agent, GeoDialogue, which is responsible for intelligence gathering and implementation of visual mediation functions. We also provide analyses of a sample dialogue to exemplify three distinct roles of visual displays in dialogue processing, and to demonstrate how GeoDialogue recognizes the intended role of visual displays and reasons about user intentions based on recognition of these roles. We conclude by a brief discussion of challenges and plans for further research.
Download
PDF Ebook Supporting Group Work in Crisis Management: Visually Mediated Human-GIS-Human Dialogue
Posted in :
