Four Stories About Computing Sciences
Have you ever wondered how your phone auto-completes your text, or how Google translates it into another language? Did you know that computers can now help you write text code and even generate images just from a few instructions?
This (and much more) is what Natural Language Processing (NLP) can do. NLP has its roots in linguistics and computer science, and it uses modern neural networks to search, classify and generate text. At Bocconi’s Department of Computing Sciences, we work to make NLP more efficient, so that it can be used on more applications and help more people in their work, stamping out hate speech and with a better understanding of minorities.
If you are interested in learning more about NLP and you like language programming, linear algebra, probability and information theory, get in touch. Watch Dirk Hovy and Debora Nozza.
Have you ever heard of genomics, computational biology or bioinformatics? Why are they good for us and how do they work?
Computer science applied to health and biology opens huge possibilities for the improvement of human health. Computational biology builds virtual models of biological systems while bioinformatics is more focused on data analysis and what we can discover by applying machine learning models to analyze large biological data sets. Genomics, the branch of biology that studies our genes, went from reading one DNA sequence in several years to being able to read an entire genome in hours.
By simulating a biological model, we can study how human diseases occur, spread and evolve over time and how to make drugs that may cure individuals with that disease.
Understanding such data, taking millions of data points and narrowing them down to a few useful results is a big challenge. To tackle this challenge, we need minds with a solid theoretical background and open to new problems. Watch Francesca Buffa and Andrea Tangherloni.
Do we understand artificial intelligence? Can we control it?
In the last 10 years artificial intelligence has entered into a new era. It is now able to analyze images and beat our best champions at everything from chess to video games. It translates texts, it helps synthesize new medicines and even writes literary essays. This technological revolution is based on machine learning: the algorithm is not programmed to recognize an image, it learns by itself from hundreds of thousands of examples to recognize it. This new artificial intelligence is also triggering rapid progress in science, from biology to genetics.
Understanding and controlling these new algorithms is the goal of our team here at Bocconi. To meet this challenge, we mobilize various scientific approaches from computer science to physics, mathematics and brain studies. Watch Marc Mezard.
Have you ever wondered how Google Maps computes the fastest way to reach a specific place? Or how the right locations for strategic facilities like hospitals or fire stations are determined?
These are problems with several possible solutions, among which the best one needs to be found. Sometimes, the number of possible solutions can be astronomically large, higher than the number of atoms in the universe. So how can an algorithm quickly find the best one?
The key to fast solutions and to fast algorithms is to find the right mathematical abstractions that can represent your data and to develop algorithms that operate on these abstractions. Such algorithms produce huge benefits for the whole society in terms of time management, costs and allocation of resources. The study of optimization algorithms is a source of fascination for those of us who have a passion for mathematics but also want to see the real-world impact of their studies. Watch Luca Trevisan and Laura Sanità.