Transcription

Machine learning and AI are systems that are able to learn from a large number of dimensions.

If you're talking about a problem that has so many different features to it that we and our limitation are unable to understand because of our ability to handle depth. When you when you train a machine with multi-dimensionality algorithms that we have today on something that is simple.

Take for example, the weather. You train it on what the weather patterns were in, say, the 2010s and you take simple features like temperature clouds, wind speed and so on. You train that network with it and then you take that model that you've trained and you apply it to today and try to predict the weather tomorrow.

It is actually able to predict the weather tomorrow much better than the weather man without knowing anything about the weather. And that in itself is very powerful and actually helping us understand that maybe we were missing the dimensionality of the world around us. Maybe that is something that the machines are going to be capable of analyzing better than we do and provide us with some insights we we never were privy to in the first place. So I think there is a lot of hype around what we see as applications that there is also the flip side to that, that in some way machines could actually turn out to be a lot more intelligent than we are in that aspect.