We teach to machines

The science of data is a hot topic nowadays, everybody knows!

The amount of measurements that our smart phone stores or has access to, can be compared with all of the science for many centuries. This allows us to start analysing people's behaviour and understand how they behave and why. Your mobile phone knows your secrets much more than anybody else, you may lie to somebody, but it is difficult to lie to your mobile phone. 

Now we need to innovate some robots to tell us what does this massive data mean! But why robots? because we cannot do it ourselves.

Analysis of such a massive data requires a lot of mathematical operations that a human being during his life cannot do, even a small portion of it (even using a calculator).

We make some machines and teach them to learn and then analyse data. We require a bunch of machines to do the analysis more efficiently, because your data keep growing as your life goes on. We (humans) learn collectively by dividing the task beween us and aggregating the final result. The new generation of computing machines such as the GPU and the CPU servers benefit this model. They  implement several tasks independently, and aggregate the final result eventually.

Our Tesla K40 is installed today at Poly, try some parallel computing tasks there.