Abstract-It can be very difficult to create software systems which capture the knowledge of an expert. It is an expensive and laborious process that often results in a suboptimal solution. This article proposes an approach which is different from 'manual' knowledge construction. The described system is relevant and usable for the end user, from the beginning of its development. It is continuously being trained by the experts while they are using it. This paper shows an example and explains how the computer learns from the experts. In order to provide a general mechanism, the principle of learning to learn is proposed and applied to the problem of handwriting recognition. The study pays attention to the development of the hypotheses that the developing system uses as it adapts to feedback it receives from its own actions on itself, on objects, on the users of the system and the reflection on this feedback. The ultimate goal is the creation of a system that learns to 'google' through handwritten documents, starting from scratch with a pile of raw images.
I. INTRODUCTIONIn the creation of systems that capture the knowledge of the users quite often the focus is on how the information should be presented and on how the creator of the system can extract the information from the users. The current study aims for a different approach. The focus is to identify mechanisms by which the system may determine itself what which information should be presented to the user in order to:• minimize the interaction (e.g., as measured in amount of mouse clicks) • maximize the information gain for the learning algorithms These two aspects go hand in hand. How to solve them is difficult and there is no abundant literature teaching us how to combine these two questions at once. The proposed methodology is based on current machine learning techniques to