Question: How do you teach computers to think like humans through code and what coding language do you use?

Oliver Gordon answered on 4 Mar 2019:

I use a language called python. It’s named after a very popular TV sketch show from a few decades ago called Monty Python’s flying circus. It’s one of the most popular languages in the world (and also absolutely free!)

As to HOW you do it, there’s a variety of ways. The most popular is to use a set of tools called tensorflow, which is made by Google.

https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.11001&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false

They also have a frontend for it called Keras, which we and most other researchers use. What you do is give the computer lots of examples to learn from. For example, if we want it to learn to tell the difference between nice/not nice cakes, we give it hundreds of thousands of examples of nice cakes, and hundreds of thousands of not nice cakes. It then “learns” to separate them, by doing a load of highly dimensional maths that I couldn’t even begin to understand, let alone explain! You then check it by giving it some cakes you know are nice/not nice, but don’t tell the computer what you think – you ask it for its opinion and see if it matches or not.

In practise, getting that data, and getting that data perfectly organised and tested and making sure your computer isn’t trying to trick you is the real challenge. There’s a reason why Google, Facebook and so on are so desperate to get as much of your data as they can (even if it means doing things that are legally questionable). Getting that data is so, so, so important – because without it you can’t do anything!