Smemo: a Multi-modal Interface Promoting Children's Creation of Personal Conversational Agents

Recent improvements in natural language processing pushed the development of a wide range of products and applications that allow natural spoken interaction for children. Previous studies explored conversational agents for children and found out that they are valuable both for playing and learning. In addition, recent researches highlighted that, when we want to develop a conversational tool for children, we must consider that they have different modes from adults and they approach conversational technologies in a different way because of their linguistic skills, their way of conducting dialogues and, more generally, their interaction attitude. In this paper, we describe Smemo, a multi-modal interface for children aged 6 to 9 years old. This tool permits to approach conversational technologies in an original and multi-modal way giving the user the possibility to customize and personalize the agent. The main goal of Smemo is to promote an innovative learning method by putting children at the center of the process.