Exploring Self-Explanation: The Human Side


Participant at the Doctoral Mentoring programm of the 10th German Conference on Multiagent System Technologies (MATES 2012)

Following the idea of ubiquitous computing, several visions were conceived and different software engineering paradigms were introduced. We claim that the agent-based approach is promising for this domain which is usually complex, distributed, and evolving. Once it comes to the development of such systems, one can easily imagine that the self-* properties are needed to induce the desired behaviour. The self-* properties describe systems that are capable to adjust their internal state as a result to exogenous and/or endogenous influences and without a centralized entity that coordinates such adaptation processes. Admittedly, this is a rather high level -- In more detail, self-* is commonly identified by self-organization, -configuration, -healing, -protection, -explanation and context-awareness. Here, we find that the access to the agent-system by human users, as part of the self-explanation capability, is often disregarded and should be explored.

Sebastian Ahrndt
Conference Paper
10th German Conference on Multiagent System Technologies, Germany, Trier