re: Explain neural networks Like I'm Five VIEW POST


Well, son, imagine you're playing Tic-Tac-Toe with one of your friends: everytime you need to put an O on the grid you do instinctively but in that short time before you actually put that O, your brain has made some calculations. Your brain said to itself: "Well, there is a X in the top-right corner! What would happen if my master put an O below it? Or shall he put an O in the left-bottom corner? Or shall my master surrender and play with that cute girl Jenny?". You don't ever ask for it, your brain enjoys help you. Well, the computer you use to see Peppa Pig on Netflix has no brain so some smart men tried to create one and make the computer smarter so they came up with "neural networks": it's a scary name but don't worry, boy. A neural network try to imitate our brains by taking a set of inputs (like the position of the Xs your friend Greg put on the grid or the Os you have already put on the grid) and it gives each input a weight (which will vary), then it tries to solve the problem by varying the weights value and passing the inputs to various layers which imitates our brain neurons. The first, second and even the third time the neural network will fail and lose the match but with time it'll become smarter and eventually the computer will become a Tic-Tac-Toe champion.

code of conduct - report abuse