Towards a Consensus on the Communication Between User Modeling Agents and Application Agents - Short Version

My presentation provides an eeort towards nding a common understanding on how user and learner modeling systems, if seen as agents, can communicate with the applications. The fundamental characteristics of the communication between application agents (Apps) and user or learner modeling systems (ULMSs) will be examined in order to draw some conclusions on what features should be considered within such a proposed communication consensus. Taking that into account I then open to discussion a proposal for a framework concerning such a communication. This framework is based on a reinterpretation of the KQML within a user and learner modeling scenario and describes the types of performatives that were considered relevant from a user and learner modeling perspective. However, due to the diversity of possible applications, ranging from Intelligent Learning Environments (ILEs) to simple information systems, some of the relevant performatives here presented may not be handled by all types of applications. Finally, I will draw some implications on how some current general user and learner modeling systems can be changed in order to convey with the proposal here presented.