Modelling Shared Knowledge and Shared Knowledge Awareness in CSCL Scenarios Through Automated Argumentation Systems

Abstract

Abstract.Over the last few years, argumentation systems have been gaining increasing importance in several areas of Artificial Intelligence, mainly as a vehicle for facilitating rationally justifiable decision making when handling incomplete and potentially inconsistent information. Argumentation provides a sound model for dialectical reasoning, which underlies discussions among students when solving tasks collaboratively in a CSCL environment. In this setting, we identify the problem of constructing Shared Knowledge and its related Shared Knowledge Awareness. While Shared Knowledge refers to the common knowledge students acquire when they work in a collaborative activity, Shared Knowledge Awareness is associated with the consciousness on the Shared Knowledge that a particular student has. This paper presents a novel approach to model Shared Knowledge construction and the associated Shared Knowledge Awareness through an automated argumentation system.

An inquiry dialogue system

“…In some situations, it may indeed be more appropriate to pool the agents' beliefs (e.g. as part of computer supported collaborative learning, [27]); however, in many real world scenarios, such as within the medical domain, there are often privacy issues that would restrict the agents from simply pooling all beliefs; what we provide here can be viewed as a mechanism for a joint directed search that ensures the agents only share beliefs that could be relevant to the topic of the dialogue. It could also be the case that the belief bases of the agents are so vast that the communication costs involved in pooling all beliefs would be prohibitive.…”