A target driven approach supporting data diversified generation in IoT applications

Smart IoT technologies have the potential to make a breakthrough in the sup-port of Cultural Heritage (CH), by providing information and communication technology to effectively enhance current models of art recreation and enjoyment. To turn such potential into reality, IoT-based technological solutions for CH should be designed by taking into account two main factors: on the one hand, they must be able to involve and attract different types of users, on the other they must avoid focusing users’ attention solely on the smartness and novelty of the supporting technologies, thus diverting them from living the experience of being in a cultural site. To this aim, endowing IoT applications with anthropic interfaces seems a promising way to explore, and most prominent among such interfaces are those based on capabilities for Natural Language Understanding and Generation. A set of linguistic resources and Natural Language Generation (NLG) techniques are needed in order to reach this goal, which we describe and characterize here through the case study of a workflow that supports the automatic generation of artworks.