One of the most rewarding parts of our relationship with dogs is how we humanize them, they seem to smile at us, speak and understand our feelings. But in the end we are just putting a human layer on an animal with a brain that is still a mystery to us. What we know very well is how to train them, we study their reaction to certain basic stimuli and take advantage of it to get them to perform a specific task. This method was set thanks to the behavioural studies developed by physiologist Ivan Pavlov. Now, more that 100 years after, his research became relevant again because we have to deal with a new kind of mysterious brain that needs to be trained: deep machine learning.
This way of programming computers differs from the traditional way because it no longer needs to by done by precisely coding every task. Instead the programmers feed data into the computer and let the machine figure out what it's being fed.
As Jason Taz from wired explains: “If you want to teach a neural network to recognize a cat, for instance, you don’t tell it to look for whiskers, ears, fur, and eyes. You simply show it thousands and thousands of photos of cats, and eventually it works things out”. The problem with this is that, since the computer does the job by itself, we can’t see how they processed the data and improve, so we need to train them like dogs.
We all evolved with a natural affinity to train and domesticate dogs, so if we treat our computers as pets we could start making amazing things with them in a completely new way and code will no longer be a barrier.
Read more at Wired. Image: Suwalls