Modern machine learning methods use computational shortcuts to handle complex models and large data sizes. This inevitably results in information loss. To confidently use such approaches a deep mathematical understanding is necessary and it is especially important to rigorously quantify the uncertainty of the procedure

We need to build equations that, starting from theory, are capable of describing phenomena of interest in the real world: just like the Nobel Prize for Physics did. And we need to do it by choosing the least complex, most transparent and therefore understandable path. Only in this way can AI truly be integrated into our lives without being viewed with suspicion

Data entry, bank reconciliations and the processing of invoices are all routine activities that are increasingly entrusted to AI technologies, something which makes data reliability crucially relevant. And so companies are now increasingly pushed to embrace forwardthinking analytics

The challenge is to make these philosophical zombies work together with humans and thus overcome the limits of both to generate a new kind of digital marketing that manages to overcome the current obsession for optimized personalized communication via AI to reconnect individual users with the collective values of a brand

A transformation in business and governance and stronger attention to fairness, transparency and privacy. Thus credit institutions are changing skin thanks to AI

Deep learning allows several applications in medicine, such as in the case of patients with breast cancer. For such systems to be effective, however, they must be safe, equitable and needsfocused