
Full Professor
DEIB - Data Science and Bioinformatics Lab (Bld.21)
March 12th, 2025 - 5.00 pm
Contact: Silvia Cascianelli
The series of weekly meetings are primarily meant for all the 'Data Science and Bioinformatics lab' members, but anyone potentially interested in the topics is more than welcome to attend.
This seminar will be held by Prof. Stefano Ceri (Full Professor, DEIB, Politecnico di Milano) on the following subject "Data Science".
In this seminar, I will give a broad vision about Data Science, as it stands in 2025, recalling some of its recent developments. I will start with the so-called “Fourth Paradigm” (2009), as framed by Jim Gray in his work with astronomers. As in many other crucial technologies (e.g. transactional systems, data cubes, …) Jim Gray was a pioneer of Data Science. I will review the key ingredients of Data Science and advocate that Data Science -- being a "discipline applied to any discipline" and having a "strong interdisciplinary flavor", is key to organizing human activities and an important asset for charactering the profile of any student/researcher/practitioner (the so-called “Phi-shaped education”).
I will also recall a debate that took place in Vienna in 2016, at Informatics Europe, where I had to contrast Moshe Vardi — he was pushing for top-down, model driven approaches while I had been invited to push bottom-up, data driven approaches. In reality, the contrast is fictious — one could start top-down, formulate hypotheses and turn them into models, and then confirm models by using data. Indeed, both model-driven and data-driven approaches co-exist in modern data science. I will also review contexts where the “big data” approach wins, using pharma companies as examples.
For what concerns education: Back in 2015, in some advanced schools (Berkeley) started merging basics in CS+Stats to build “Introduction to Data Science” courses dedicated to the early phases of education (undergraduate level); then, several Master Programs in Data Science were created (I will describe the IACS Master of Science at Harvard), as well as PhD Programs (including Polimi’s “Data Analytics and Decision Support" (DADS), attended by several students of our group). As of today (March 2025), a big push forward comes from Stanford University with the opening of Computing&DataScience building (https://engineering.stanford.edu/get-involved/support-engineering/funding-initiatives/computing-and-data-science-coda) - I will quote the impressive endorsement of the initiative of Stanford’s President, of the Dean of Engineering, and of the head of the Data Science Initiative.
In this talk, I will not talk about my research; by looking at my Google Scholar listings in years 2024-25, you may see a variety of topics, ranging from foundational research on some data science technologies up to applications to viruses (recent work on Avian Flue), law (recent work on modeling/implementing the graph of Italian laws), LLM (recent work on building a conceptual map), finance (company control problem), judicial system (duration of trials), drug repurposing, and other. In my view, Data Science is a “conceptual framework” and “mental mindset”, fundamental in approaching problems of arbitrary nature.