Computing Sciences Have a Home Here

AT BOCCONI THERE'S A NEW DEPARTMENT THAT DOES RESEARCH AT THE INTERSECTION OF COMPUTER SCIENCE, THEORETICAL PHYSICS, APPLIED MATHEMATICS, ARTIFICIAL INTELLIGENCE, MACHINE LEARNING, AND COMPUTATIONAL NEUROSCIENCE. TALENTS FROM ALL OVER THE WORLD HAVE COME TO WORK HERE AND RICCARDO ZECCHINA IS AT THE HELM

Work is in full swing on the third floor of the Via Roentgen building. After moving boxes and setting up new offices, the staff of the newborn Department of Computing Sciences is gathered here, in a structure that combines teaching and research at the crossing of computer science, theoretical physics, applied mathematics, artificial intelligence, machine learning and computational neuroscience. Leading the team, which in its initial roster has 5 full professors, 2 associate professors, 11 assistant professors and 5 lecturers, is Riccardo Zecchina, an engineer who graduated from the Polytechnic of Turin, currently full professor of theoretical physics and Bocconi Vodafone Chair in Machine Learning and Data Science.
 
What are the objectives of such a department in a social sciences university?
Bocconi has been expanding towards STEM subjects for some time. It is part of the University’s strategic plan and very significant skills have already developed regarding the use of these "hard" sciences in the social and economic fields. The convergence between economics, computational sciences and social sciences is a fact occurring in today’s context, and having acknowledged this, Bocconi seeks to immediately place itself at the academic forefront. The first objective is therefore to do research at the highest possible level and, at the same time, to interact with other disciplines. It is from the interaction with economists that new knowledge will arise: our task is to create the conditions for the interaction between data science and social science to be as fruitful as possible.
Which research areas is the new department bound to explore?
Computing sciences encompass different fields, from computer science to modeling science, computational neuroscience, and computational social science. These four pillars converge in machine learning and artificial intelligence. Within this framework there are topics such as algorithm theory, optimization, natural language processing, game theory, computational aspects of data privacy and cybersecurity, and quantum computing. There is also the area of digital humanities, where we will do applied research on the possible uses of AI in various branches of economics, from accounting to finance. Finally, we will try to carry out research in the field of life sciences related to computational biomedicine and image processing.
Which talented resources have you managed to attract to the new structure?
The first arrival was from Berkeley, Luca Trevisan, a theoretical computer scientist of absolute excellence, a reference for the global community. Alon Rosen instead comes from the Weizmann Institute for Sciences in Israel, and is one of the leading experts in cryptography and security. We have with us Marc Mézard, until recently director of the ENS in Paris, who will give an essential contribution on machine learning, while Francesca Buffa comes from Oxford and deals with algorithmic learning in computational medicine and in particular with respect to cell networks involved in the regulation of tumors. Among younger researchers, I would like to mention Dirk Hovy, already at Bocconi, an expert in natural language processing and Laura Sanità who will join us in September, specialized in optimization and operational research. Almost all of them have ERC grants, a certification of international research excellence. There are also many young assistant professors in whom we place great trust and for whom we harbor great expectations.
What contribution will you make to teaching?
We would like to help insert more computational knowledge into all degree programs, by hybridizing university training and thus responding to the very strong demand from the professional job market. We have already given birth to a three-year undergraduate degree, the BSc in Mathematical and Computing Sciences for AI, which is enjoying great success and is a purely STEM course, i.e. it is classified as an applied mathematics program. It is an interdisciplinary path that provides students with a very solid methodological background but also gives access to more traditional Bocconi graduate programs.
Considering the appeal for young people pf the disciplines you teach, you will soon be in great demand. Aren't you afraid of an overbooking effect?
Already now we have plenty of requests and probably some of our young researcher have taken on more commitments on the supervision of theses than they should have, but everything will be brought up to speed. However, we must clarify to the students that we do not teach technology, we describe a series of conceptual, methodological tools through which they can exploit the digital technologies in every field. Our aim is to create a new kind of “Bocconian”, who knows how to create models, interpret tools and create new ones to stay up to date. It is not a technology transfer, but an understanding of conceptual methods that requires a lot of student effort and gives long-term benefits.
Have you also thought of initiatives open to external users?
There is a huge problem of continuing education in companies for many of our subjects, but it is still premature to think about it. Today we must aim to bring the best scholars to Milan for research and teaching. In addition, together with three other Milanese universities – Milan Polytechnic, U of Milan and U of Milan Bicocca – we have created a joint research center associated with the European Laboratory for Learning and Intelligent Systems (ELLIS) which is developing many interactions with the industrial sector. I believe it is an important social message because it means that there is attractiveness in sectors such as information technology, cybersecurity or machine learning, and this is important for the entire Italian social and economic fabric. In short, we are training many young people who will be precious resources for companies.
 

Riccardo Zecchina, Professor of Theoretical Physics, holds the Vodafone Chair in Machine Learning and Data Science at Bocconi and is the Director of the BSc in Computing and Mathematical Sciences for AI. His current research concerns the interface between statistical physics, computer science, machine learning and computational neuroscience. In 2016, he received (with M. Mézard and G. Parisi) the Lars Onsager Prize in Theoretical Statistical Physics from the American Physical Society. Previously (2011-2015) he was an Advanced Grantee of the European Research Council (ERC). A few weeks ago, he was appointed head of the new Department of Computing Sciences. “For me, whose primary interest is research, accepting to dedicate myself to the creation of this new structure was a great time investment”, says the professor. “This is why I was really impressed by the reactivity and spirit of collaboration that I found at Bocconi. The ability to accept innovation is far from obvious. Colleagues have demonstrated an intellectual openness that confirms the high level of intellectual pursuit that characterizes our university. Our new research structure has all the features to quickly gain international visibility.”

by Emanuele Elli