Expert critiquing systems are a type of humancomputercollaborative system in which a computer agent presents reasoned opinions about a humanagent's problem-solving process for a given task. Thechallenge in such systemsis to providetimelycritiques relevant to the user's focus of attention. A problem with manyexpert critiquing systemsis that their critiques are not alwaystimely or relevant; consequentlysuch systems interfere with problem-solvingrather than provide assistance. Theproblemarises in part frominsufficient representations of the human's problem-solvingprocesses. In this paper, wediscuss the flexible use of an agent modelbased on a task-decomposition hierarchy of humanexperts in a critiquing system called SEDAR.The modeldiffers fromprevious research efforts in three ways:1) the structure of the model,2) the function the modelperforms in the expert critiquing system, and 3) the influence of the model on communicationbetween the computerand humanagents. A prototyp...
Michael Chin-Ming Fu, Caroline C. Hayes