Microsoft to work closely with Intel on Windows ML

[DanceswithUnix]

Don't get me wrong i agree with you that computers are doing things that was unheard of yesteryear, however just because they're more advanced doesn't mean they're AI, they're just doing what they've been programed to do and they do it better because they can access and process more knowledge than computers of the past, however that's been the general direction of travel for computers since they were invented.

Take the fingerprint or voice recognition on your phone, neither are AI IMO as they're just following rules that have been programed into them to do certain things when it comes across certain identified conditions, change the conditions and they wouldn't work.

This is not about better programs, this is about teaching computers by example which allows them to perform tasks that the programmer might not even understand.

For example, you can program a computer to recognise cat videos by describing in code what cats look like. That is the old way, using things like the amount of jaggedness of a cat face in this range, a certain aspect ratio which you have calculated.

The modern way is to choose a neural net configuration, and feed it videos telling it which ones contain cats and which ones don't. Get it right, and you can feed it any video and it will identify the cat ones, but the killer here is that at no point has the programmer described a cat in any way to the computer. The machine has worked out what a cat looks like. The machine has learnt.

If you don't find it at least slightly scary, then you haven't understood the technology. We are still miles off general intelligence so no need to build that panic bunker in the garden just yet, but it is still a big step change in how things work.

Edit: Better example, feed a computer medical images and tell it what ailments the patients have. The computer starts to predict from new images what ailments new patients have. Researchers then have to reverse engineer their own system to work out what features of the image the system is using to make the prediction because they want to know what it is looking for because the researchers didn't even know you could make the prediction.

[Corky34]

I think you're getting confused with the surface tasks and the methods used to execute them. Take voice control - on modern smartphones it's not following pre-programmed recognisable commands. It's doing full Natural Language Processing to derive meaning, then triggering tasks in response to that processing. NLP is an incredibly complex area of computer science and for Siri/OK Google to work they way they do without any pre-programming by the specific user is definitely down to the use of AI and ML. The fact that on the surface it looks like you're simply triggering programmed responses by using command phrases is testimony to just how good/quick the underlying techniques are nowadays.

I wasn't getting confused, i was trying to illustrate how so called AI or ML are just following rules that have been programed into them and simply making statistical judgments based on probabilities, in essence their making best guesses based on previous knowledge of data inputs and if you changed the type of data input from something like an image to sound (fingerprint or voice) that it wouldn't work because its only been programed to make those statistical judgments on specific types of data (knowledge).

The modern way is to choose a neural net configuration, and feed it videos telling it which ones contain cats and which ones don't. Get it right, and you can feed it any video and it will identify the cat ones, but the killer here is that at no point has the programmer described a cat in any way to the computer. The machine has worked out what a cat looks like. The machine has learnt.

What I'm saying is that yes its learnt but how it learns is still defined by programing it, it's simply making statistical judgments based on particular predefined parameters.

[badass]

I wasn't getting confused, i was trying to illustrate how so called AI or ML are just following rules that have been programed into them and simply making statistical judgments based on probabilities, in essence their making best guesses based on previous knowledge of data inputs and if you changed the type of data input from something like an image to sound (fingerprint or voice) that it wouldn't work because its only been programed to make those statistical judgments on specific types of data (knowledge).



What I'm saying is that yes its learnt but how it learns is still defined by programing it, it's simply making statistical judgments based on particular predefined parameters.

You seem to be inferring that AI is identical to ML. What you describe is ML. ML is one form of AI. There are other forms of AI that are not ML.