| < Day Day Up > |
|
Some researchers (Mizoguchi et al., 1988; Kono et al., 1992; Giangrandi & Tasso, 1995) applied truth (or reason) maintenance systems (TMSs) (Doyle, 1979; DeKleer, 1986) to overcome conflicts between new and old knowledge. The TMS identifies the conflicts, and some domain-specific reasoning system will resolve them. Huang and McCalla (1992) and Huang (1994) developed “Logic of Attention,” a variant of the TMS that focuses on the parts of the student model and instructional planner that are relevant to the current subgoals. Bruff and Williams (2000) proposed an architecture in which the problem of conflicting information is resolved using methods based on the AGM paradigm for belief revision (Alchourron et al., 1985). Bruff and Williams used possibility theory (Dubois, 1992) to address the problems of uncertain information, nonmonotonic reasoning, and default logic (Reiter, 1980); the formalism (Antoniou, 1996) to process incomplete information; and Theory Extraction for fusion. Knowledge granularity has been widely discussed (e.g., McCalla & Greer, 1994). Levels of granularity fit naturally into the agent architecture and can be used to help the agent choose an appropriate plan.
| < Day Day Up > |
|