An interactive environment for dynamic control of machine learning systems
Engelbert Mephu Nguifo · International Conference on Human-Computer Interaction · 1997
This paper describes an interactive model that allows a dynamic control of machine learning (ML) systems. The model is based on the design of ML systems. It integrates different tools with various purposes in an open environment in order to facilitate the interaction between the ML system and another agent which is often an expert or a user. This integration is done in such a way that every change of data in one of the tools is automatically taken into account and interpreted by the other tools. Such dynamic process improves the interaction with the ML system, and consequently the interpretation of ML results. A prototype of such model is described.