Understanding what happens inside artificial intelligence is one of the main tasks of modern times. Although developers can observe the data fed into the model and the answers it generates, they are not always able to explain how the system arrived at a particular conclusion. However, new research suggests that this may be solved with a new theory.
The study involved Roberto "Pena" Spinelli, who is a physicist from the University of São Paulo and has a specialization in machine learning from Stanford University.