Implementation, interoperability, and scaling Working group ICT Workshop on Intelligent Lifelong Learning Companions
Maria Bauer · 2008
Summary The realities of building an intelligent lifelong learning companion could very well be the most challenging hurdle of all. The huge number of systems, courses, instructional materials, and social influences that bombard a learner are both motivation for building a learning companion in the first place, but also intensely challenging to manage automatically. Intelligent tutoring systems have made great strides in isolated learning contexts, including supporting learning over semester‐long periods in a single subject. However, the leap to supporting multiple subjects and integrating modeling techniques across domains and over time (both in modeling expert knowledge and student learning) represent major hurdles. In addition, the reasoning required by a lifelong learning companion, as Minsky et. Al (2004) have discussed, imposes further demands on the representation of knowledge and scaling up of any kind of tutoring or assistant technology.