Gabriele Musillo

Gabriele serves as dMetrics Director of Research. In this role, he initiates and executes core research tasks in statistical NLP. Gabriele holds a PhD in Computer Science from the University of Geneva, where he also completed his undergraduate studies in logic and linguistics, and was a postdoctoral research associate at FBK-irst. Gabriele has also served as Senior Lecturer at the University of Geneva, teaching the applications of deep learning to NLP. His research work has focused on data-driven statistical techniques for learning deep relationships between different levels of representation of linguistic data. In particular, he codeveloped the first recurrent neural network to jointly parse syntactic and semantic structures. His work on structure prediction models integrating latent states and shallow semantic information has been published in the most highly regarded academic conference proceedings and journals in the field. His current research focuses on integrating interactive machine learning and representation learning, to make the process of teaching machines near real-time, collaborative, effective, and accessible to any end user.