Data-driven strategies: data as the basis of both the information system and the training system was the subject of a meeting organized by the company Denodo and which featured a very varied audience, made up of academics, analysts and business leaders. More in detail, stimulated by the questions of Andrea Zinno, data evangelist for Italy by Denodohave provided their opinions Silvana CastanoProfessor of Computer Science dell’University of Milan; Raniero RomagnoliChief Technology Officer di Almawave; Daniel BobbaSenior Partner at Deloitte Consulting, Big Data & IDO Leader; Richard TorloneUniversity Professor ofRoma Tre University; Gabriele Obino, Regional Vice President & General Manager Southern Europe & Middle East of Denodo. Here is what emerged.
– It is undeniable that data is taking on an increasingly important role. How is the university world perceiving this acceleration?
Silvana Castano – The educational offer of our department is quite varied. We have four undergraduate courses: information technology, information technology for digital communication, music information technology and security of computer systems and networks. To these are added two master’s degrees in computer science and cyber security. For three years, we have enriched the training offer, which was very geared towards computer science, with the new degrees in data science for economics and artificial intelligence, which see the inter-university partnership with Pavia and Bicocca. There are also a bachelor’s degree and two master’s degrees in artificial intelligence. The last one, the more human centric one, starts this year. It is, therefore, evident how the theme of data is central to the training course. In fact, it always has been, but it has profoundly evolved to where we have entered curriculum specifici in data science given the importance of data in business and in the various application fields, such as the medical sector for example.
Then there are the issues of governance and of cybersecurity. What we have experienced directly in the new degrees, and also by colleagues, who have never been asked of us like now in the past data management and coding courses. There really is a perception of the importance of data not only in those involved in information technology but also in other fields.
– How do the needs of those who deal with data every day change?
Gabriele Obino – The data has always been there, but there has been an evolution that has come in waves. First it was the databases, then the content tools for doing data analysis, then the data warehouse and then the data lake. The point is that data grows like the entropy of the universe at an astonishing rate. However, the data in itself has no value. If you can’t connect it to the context in which you want to use it, it’s useless.
Data management arises from the fact that we are drowning in data and there is an urgent need to extract value. And even if artificial intelligence helps, the fact remains that people need to understand what the meaning of the data is. And to understand it maybe it’s time to stop copying data indiscriminately from on prem to cloud, from one cloud to another cloud or from one application to another. We need to contextualize the meaning, focus on metadata which help to give meaning to the data, to build a context around which to truly understand what is the phenomenon being analysed. What is missing, in my opinion, is to bring a very high capacity of interpretation because this then helps companies and business a lot whatever the context.
– You are big data users. What care and caution do you need to take when using data?
Raniero Romagnoli – The ecosystem is complex because the legislation is not one, but many. And they continue to increase. More and more it also tries to standardize artificial intelligence. In the case of using data to train a model in the compliance with the GDPR we have quite tight boundaries that force us to use very broad precautions.
Then there are techniques, especially within artificial intelligence, that allow for develop further algorithms or approaches to segregate, to disguise to do federal or deferred trainings. These algorithmic structures support us in doing the trainings. The data that we then use if they remain at the customer’s home, and we hold responsibility for the treatment at the customer’s home, everything is simpler. However, if we put the data on a platform is quite different: we cannot use the customer’s data unless the customer gives the data to train the model with all the terms we know about privacy. So it’s a very complex process, there is no clarity on the AI Act nor on the direction it takes, although some pillars are very clear on what Europe would like to do with artificial intelligence.
– According to Deloitte’s point of view, what is the position of companies towards artificial intelligence?
Daniel Bobba – Il data valuethat is the fact of being able to be credible and have information data reliable, is critical. L’artificial intelligence is the tip of the iceberg that implies the corporate understanding of a culture that leads to reading the data not as something derivative, but as something that characterizes a real fact. In the ideal world, a business process is represented linearly with a system that rests on a database.
Companies are hungry for data and they are happy to know that universities will churn out large numbers of data management experts and there will be no more dramatic difficulty in finding new people. We have an incredible demand for experts: by now all cross sector companies are undertaking data driven transformation paths. There are companies that have done this by nature, such as those in the world of financial services because they must comply with precise regulations. There are industrial or financial sectors that have been using artificial intelligence for decades, there are others that are facing it now.
I think that Italy is not badly positioned in Europe. Perhaps compared to the American world we suffer from the fact that we we don’t have big IT companies, but many of the excellences in artificial intelligence are Italian startups. But an important thing arises ethical theme which leads to asking how a decision was made. In addition, as mentioned, to the fact of the need to know if information is reliable in the way it has been defined.
– How much has the collaboration between university and company changed over time?
Richard Torlone – I have to dispel a myth that of a university perched on theoretical aspects that does not look at the business world. This is absolutely not the case. And it hasn’t been for many years, but there is a very strong focus on what is happening in the business worldin the markets and on what companies expect from training. The numerous basic courses are accompanied by a very dynamic offer, especially in the master’s degree. I believe that in the last 10 years the courses that we have offered in the master’s degree have changed a lot and every year we introduce new courses that naturally also start from our skills, our knowledge and they do what the corporate world expects. For example, we have many active initiatives that support youth entrepreneurship and all that leads to world of startups. We also have an incubator within our department. We also organize meetings between students and companies.
Gabriele Obino – Denodo was born in the university environmentat the University of La Coruña, from an idea of professor Angel Viña who imagined a virtualization-based data integration strategy, a strategy where users could integrate data without replicating it. Over time we have always maintained a direct contact with the academic world so much so that he launched the Denodo Academy Program with which it tries to create a sort of link between university and company as soon as possible by offering interested students our technology and our expertise.
– What is the relationship between data and sustainability?
Gabriele Obino – It is a topic of great actuality and relevance. Data and sustainability don’t seem related, but that’s absolutely not the case. It is important to be able to produce sustainably. Since it is a new theme it is cross discipline. To have a sustainable business, companies bring together data from multiple systems and, therefore, first of all they need to understand what data they need. Information arises from the context within which the data takes on meaning. Instead of talking about where data is stored we should talk about what information is neededthus changing the approach from bottom-up to top-down.
Andrea Zinno – Then there is also a less fascinating aspect which concerns the way in which the data life cycle is managed. We are used to replicating filesso data proliferates and they are stored on systems that consume energybut usually you don’t think about it. Hence, the fact that they were born architectures that allow for take the data where it is without having to replicate it each time brings an overall efficiency and also has a impact on resource consumption (and therefore on the environment). Indeed, every time a data is movedbeyond the security and governance implications, and consumes energy.
Raniero Romagnoli – Speaking of ESG, there is a regulation which has already been obliging since 2023 large companies, and from 2026 also small and medium-sized enterprises, to make sustainability reports. Unfortunately, there are still too many frameworks and there isn’t a single one for this indication.
Then to be emphasized the energy consumption that requires artificial intelligence. Only for the last ChatGPT training three 4.5 million dollars were spent and the operation required 1 million hours of GPU usage. This is equivalent to tens of tons of CO2 released into the environment. And even when you ask questions to ChatGPT you have important consumption. whereby, 3in many cases we can refer to classic models, such as linear regression, which are just as effective but have much lower consumption.
Silvana Castano – In the academic field, and more generally in the Public Administration, we have the gender budget. It is a document in which we must also address the issues of sustainabilitydell’inclusiveness and ofgender equality, validating the policies and actions put in place to also guarantee the well-being of the community. About the didacticsince last year we have a new master’s degree courseThe sustainability of natural resources” which touches on all of these issues. When there are teachings with a syllabus that talks about sustainability, inclusion or low environmental impact, these themes become an important educational pathwhich can also be interesting for a computer scientist.