Focus: Currents

From data flows to shared knowledge

Flussi di dati
by Benedetta Ciarmoli

Every day, we generate enormous quantities of data that flow through digital networks and devices. How can these streams of information be transformed into knowledge that benefits society? The answer depends not only on our ability to analyse them, but also on sharing them and bringing different disciplines together, as Valerio Grossi of the CNR’s Institute of Information Science and Technologies in Pisa explains

Published on

“My story begins with stream mining”, says Valerio Grossi of the CNR’s Institute of Information Science and Technologies (ISTI) in Pisa. “Stream mining is a technique for analysing data as they arrive, adapting to changes in the phenomena being observed. Think, for example, of a system that recommends products to buy online: if it relied solely on purchases made in previous years, it might suggest items that no longer reflect consumers’ interests. With stream mining, however, the system analyses incoming data and updates its predictions, adapting to changing preferences. What changes, therefore, is not just the amount of information available, but also the knowledge we can derive from it”.

However, the possibility of creating new knowledge does not necessarily depend on collecting new data. Sometimes, all it takes is learning to look at existing data from a different perspective. The researcher cites DBLP, a database originally created to catalogue scientific publications and make them accessible. “The same information can be used to study researcher mobility by reconstructing career paths and tracking movements between countries. This is perhaps the most interesting part of my work: using data sources that were created for entirely different purposes”, he continues. A body of information intended to answer one question can therefore become the starting point for an entirely different line of research.

It is precisely this potential that highlights the value of sharing. Grossi recalls how, even within the CNR laboratories in Pisa, there were datasets that other research groups did not even know existed. On a European scale, this challenge becomes even more evident. The need to connect resources and expertise led to the creation of SoBigData, the European research infrastructure coordinated by CNR-ISTI and dedicated to big data analytics and social mining, namely the extraction of knowledge from data describing social behaviour.

Correnti informatiche

Alongside sharing, a multidisciplinary approach is another key element. “When you bring together different kinds of expertise, you also bring your own experience, and that includes data. In a project examining the impact of artificial intelligence on society, for example, involving psychologists allowed us to complement computational indicators with new perspectives on how to interpret the findings”, Grossi adds.

Bringing disciplines together does more than enrich the available information: it changes the questions we can ask and how we interpret the answers. Sharing, however, also faces obstacles. Data are a scientific and economic resource, and those who hold them may be inclined to protect them, particularly when their use offers a competitive advantage. It therefore becomes essential to strike a balance between openness, personal data protection and the public interest. Public research can play an important role in maintaining this balance by developing tools and services that are not driven solely by commercial considerations. There is also a more fundamental limitation. “Data cannot tell us everything”,’ the researcher reminds us. “Recorded information is necessarily discrete, whereas human behaviour is continuous and complex. Even when data streams allow observations at ever shorter intervals, no set of measurements can fully capture human experience”.

The challenge, then, is not simply to collect and analyse information ever more quickly, but to learn to use it thoughtfully and collaboratively. Looking ahead, Grossi highlights the need for humans and artificial intelligence to evolve together, with researchers and society also learning to understand and govern tools that play an increasingly prominent role in everyday life. “In this sense, the real strength of digital currents lies not only in the speed at which they carry information, but also in their ability to connect data, knowledge and people. These connections, too, can help turn data into benefits for society as a whole”, the expert concludes.

Source: Valerio Grossi, Institute of Information Science and Technologies,  valerio.grossi@isti.cnr.it