Supporting Knowledge Modelling by Multi-modal Learning: Defining the Requirements

A persistent problem faced in modelling medical knowledge is the required combination of skills from the fields of computer science and medicine. The expertise of experts both in medicine and in computer science has many dimensions which makes competence deficits hard to target. Our approach is to provide the required training for the modelling job integrated into the editing environment, providing training on the job and mutual support by modellers with complementary expertise. In this work, we analyse the requirements of an editing environment that blends various forms of learning with the editing process itself. In particular, we need to combine (1) static online information of various granularity, (2) support for asynchronous interaction with advisors, (3) tracing of user actions and pattern matching of similar use cases in the past to support the learners memory, and (4) automating parts of the modelling process to produce suggestions automatically where possible. We are currently integrating the above into a system which is expected to provide a productive environment for experienced users and a smooth and efficient learning experience in those subdomains of knowledge where the user is less seasoned.