Why Unsupervised Learning Is the Best Approach for AI Applications

Most AI applications nowadays are based on supervised learning which means that you have data from the past but these data have been labeled or annotated with information of interest. In unsupervised learning your task is normally not so much on making predictions but on trying to infer or to detect some underlying structure or patterns within that data. We don't speak about training in fact because we are not training on any power state that we just discover some patterns or some structures from the past data that you have collected. So this was the main reason and then more scientific reason is that we wanted to see indeed if there was some similarity between those antitrust cases now it's clear that if you want to train the data you