Artificial Intelligence
UC -- A Progress Report
User interfaces and help systems: from helplessness to intelligent assistance
Artificial Intelligence Review
Hi-index | 0.00 |
This paper deals with the problem of assigning meaningto the interaction of a user with a command-drivensystem such as UNIX. This research is part of theconstruction of an active intelligent help system thatmonitors users in order to offer spontaneous help whenthey are facing problems. In order to ensure this, thehelp system must build and maintain a model of theuser. We describe a mechanism that is used by the usermodelling component of such a help system. Thismechanism makes explicit assumptions about the userwhich account for different hypotheses about what theuser is actually thinking at every stage of theinteraction. The consistency of these assumptions ismanaged by an Assumption-based Truth MaintenanceSystem (ATMS). The selection between a number ofdifferent user models is based on the informationwhich is extracted from the feedback that UNIXprovides in response to user actions.