Target-Monitoring Learning Companion Design

Abstract

Under the big wave of artificial intelligence, one kind of critical technology is chatbot, whose virtual identity can play an important role in a virtual community. By proposing the system model of an Internet community and its chatbots, this paper designs target-monitoring (TM) intelligent agent and applies in real learning community. For a learning community platform, the system outputs can be assessment outcomes and learning attitudes, which can be transformed to linguistic variables by fuzzification. Moreover, a learning chatbot is designed with three-layered structure: fuzzified status, decision table and generated sentence. For learning, certain teaching strategies can be coded in the decision table to make chatbot generate advice sentences to learners. To prove the idea, a chatbot structure for learning monitoring is implemented within LINE community, and a real learning experiment is carried out. Moreover, such chatbot and its community system structure can be not only used in learning; an extended experiment is planned to construct a TM agent for operation management.