This exciting project tries to investigate language compositionality (from the word to sentence levels) through human-robot interaction with a view to making a robot able to understand human instructions. This requires inferring the latent syntactic structure of language through perceptual information (e.g., visual and/or haptic perception) and defining its mechanism of compositional development. This project will set the basis for learning the meaning of words and their syntactic categories (e.g., nouns, verbs) through perception in the surroundings, and learning language grammar - through an unsupervised approach - so as to understand phrases and sentences (more details and research questions are available in the link below).