Expert Interview: Rocco Tripodi
Expert interview to our Rocco Tripodi
1- What is your background and role in Polifonia?
My background is in computational linguistics and lexical semantics. Currently, I am an Assistant Professor at the University of Bologna. Previously, I worked as postdoctoral researcher within the Sapienza NLP group at Sapienza University of Rome and the European Centre for Living Technology in Venice.
I like to play with words. My main interests are in fact in game-theoretic models for natural language processing. I am also particularly interested in the design, learning, and evolution of linguistic communication systems.
In Polifonia I will work on extracting musical information from text. I will try to make sense out of tons of documents stored in different digital libraries. I will automatically scan textual resources, selecting pieces of information, and trying to combine them in a large mosaic. This will allow us to set up a computational resource for browsing historical information about music and to connect it with its socio-cultural context.
2- What is your motivation for conducting research in this field?
I am particularly interested in semantics. I like to investigate how words work and what they can tell us about the society that introduced or modified them. Since my childhood, I liked how words are combined to create jokes. I must thank my grandmother for this. Since she used to entertain me and my sister with jokes and anecdotes.
3- What interests you the most in the Polifonia project?
I like the diachronic, multimodal, and multilingual components of the Polifonia project. These three elements are fundamental for the development of innovative technologies in Artificial Intelligence. Nowadays, we are seeing advances that enable a machine to create music or art, but to what extent can this machine be creative? Besides developing tools that can be used for the preservation of cultural heritage, one of the key challenges for Polifonia is to create data that can be used to train and control neural networks. I am referring to projects such as MuseNet. Instead of training networks only with raw sounds, I think that supplying them with more contextual or multimodal information can give more “awareness” to the machine. This will also be more controllable.
3- Do you play an instrument? What is the place of music in your life?
I started playing guitar when I was a teenager but, unfortunately, I am not practicing much, recently. I would like to give more time to it and to start playing drums.
Music plays a central role in my life. My memories are always connected to songs. There is the summer of Manu Chao, the afternoon of Nirvana, the night of Mingus plays piano, the morning of Satie, the spring of Pat Metheny and Charlie Haden, and so on…