יובל שחר

In this paper, we compare the methods for solving two very different tasks that employ explicit models of context: the abstraction of time-oriented data and retrieval of relevant medical text documents. We define formally and compare the task-specific ontologies for each problem-solving method. The mappings between these two problem-solving frameworks provide a basis for defining a more abstract task - context-sensitive pattern matching. We discuss the implications for reuse of the knowledge represented in the ontologies and inference structures, and we suggest similar tasks that might employ the generalized context-sensitive pattern-matching task.