An important requisite for an efficient interface is the ability to learn to recognize new objects, their function and to understand the actions of the person with whom it interacts and to anticipate an adequate response. The focus of this thesis is build an user interface with the ability to develop these skills in interaction with an instructor and/or with another person. On the methodological level the work will be based on learning and adaptation in a dynamic-neural-field-architecture, for human-computer interaction, that integrates known neuro-cognitive mechanisms supporting joint action in humans. As an interaction scenario we choose a very simple scenario, where the interface could learn the colour name and fruit name of the three fruits of that scenario.