A lot of gibberish is being bantered about these days about machine learning.
Ah, people: machines cannot learn! They can only accept data and put it someplace. That is NOT learning.
It is just one of the many things wrong with today’s state of the ‘art’ of artificial intelligence, besides that whole ‘intelligence’ thing.
To know is to learn. To file away is to ignore what is kept. No learning in AI. Yet.
So what is the difference between learning something and simply filing something away for later?
To Know, one must be able to know that one knows. Or one will not know, that one knows. Current AI cannot know.
Knowing is embracing knowledge and finding it fit other knowledge and causes the emergence of the ‘common sense’ of the topic. AI does not
relate data it only weighs connections. Nothing emerges from that other than predetermined output.
If new knowledge does not fit current knowledge it is automatically rejected as not relatable. Do that enough and it works its way into the
emerging knowledge whole.
Knowledge makes up the parts of the brain’s function, processed as memory. The whole that comes from that knowledge is what allows knowing.
A person can sit in a lecture hall for months on end and walk out having heard every single thing said, and not remember a thing. Really.
Learning is when you do something with knowledge. Whether that is embrace the newness or match to your past, you will not learn unless you connect the input to something you already know.
What you already know about learning is:
Definition of learning (Merriam Webster )[1]
1 : the act or experience of one that learns a computer program that makes learning fun
2 : knowledge or skill acquired by instruction or study people of good education and considerable learning
3 : modification of a behavioral tendency by experience (such as exposure to conditioning)
Scratch #1! Both #2 and #3 are the same thing. Acquiring knowledge is not learning. Connecting it to something you already know IS learning. But if that connection is not controlling, you didn’t learn a thing.
It is why it is so easy to brain-wash an entire country. Say something nobody knows anything about and everyone will reject what you say. Those threatened by it will reject you as well. Say something everybody has heard about and now you have a listening audience, ready to be primed with more elementary parts of a larger goal, one at a time imposed upon the memory. Eventually those people won’t even know they are no longer the same people.
Liberals have mastered this art.
Republicans talk between each other about the stupid things liberals say and do. Liberals just say and do them. Republicans therefore, waste their required response by wondering about the sheer idiocy of the issue.
Take Joe Biden. (Resisting Henny Youngman.) Just yesterday the man stated that some 700 million plus people would be back to work under his plan. The country only has some 320 million or so people in it. But it doesn’t matter. He also made ridiculous statements on other topics. A Gaffe machine? Nope! The democrat press is not acting like the conservative press is. While Republicans laugh and wonder how the man can breath on his own, Democrats just watch low information voters pick up on parts of speeches and the gaffes do not matter. In fact, the gaffes keep the non-democrat press occupied while the real direction is being taken.
What matters is the public is being taught (learning) things that are not only untrue, but in many cases are seditious.
And learning none of the gaffes matter. There is no such thing as a true or false gene. You are not born with a sense of right and wrong*. You are born with a blank slate. Your upbringing and early life determine your comprehension of right and wrong. What is right to one, could be wrong to another and the other way around as well. Learning requires prior learning.
If learning did not require prior learning, babies could become instant professors and AI could really have anything to do with intelligence. A collection of data stored in a computer with no connection to any other collected data without a program to make that connection puts the donkey behind the Cullinan. A mixed up creature is trying to push what it cannot hope to comprehend. But if it takes years, and it is offered in small steps and things are built on other things: a personality can emerge that is the whole of the data acquired.
Knowledge is not input.
Knowledge is not a memory.
Knowledge is the connection between memories that emerges as a whole.
Intelligence is not the collection of knowledge or the process of assembling it: it is what does work with it.
Stephen Hawking spoke of the illusion of knowledge. He was right in that instance.
STOP calling it machine learning. There has only been one and it is in a crate.
[1] https://www.merriam-webster.com/dictionary/learning
* You are born at a specific frequency which remains you. If that frequency is odd it can never collide with itself in the brain and cause confusion. If that frequency is even, it will collide with itself and in some cases amplify a memory while in others completely cancel a memory. The latter is the source of pure evil human behavior.