That intelligence is in the stride.
In essence, it is mathematics that tries to mimic a trait of man (which is said to be intelligent).
We think of recognizing 10 species of animals on photos.The algorithm is trained to be able to recognize a cat in photos with a certain degree of accuracy.
There are obviously other methods where the machine learns in other ways.Then just based on references of correct and wrong data.
Algrotimes, calculations, statistics and desired outcomes are important concepts.Intelligence unfortunately not. Aided by ridiculous marketing like Sophia the robot becomes the first recognized robot inhabitant of a country, that is Monthy Python absurd. People think directly about robots with feelings. Robots that inefficient people find and thus eradicate. Robots that will make tens of thousands of jobs disappear next year. Quite tiring all for Salix in Pakskes.
What is significant is that today hardware proves fast enough to perform certain simple trained models real-time.That the cloud and datacenters now install Gpus and FPGA to accelerate AI training. That AI is no longer the domain of academics but also of hardware and software engineers. We can speak of a real start, but the curve is exponential. We are in a very slow rising beginning.
That is just about what it stands for.
Mainly the “intelligence” part.
What we now see as artificial intelligence comes down to a statistical analysis in which we learn a program or algorithm to predict, based on data in the past, a given in the future.
This is not an intelligence that can reason and predicate based on logic or logical reasoning.
So an artificial intelligence that is going to take over the world is not there for the time being:-).
That the term “intelligence” expresses something we understand.
Measuring, let alone understanding, of intelligence is something we still have no grasp on.The remarks that AI so far are “just statistical tricks”, the plank is completely wrong. Look at the play style of Alpha Go or Open AI at Dota 2. That is not a trick, that is truly an understanding of the game at a level that no man has previously been able to achieve. That is intelligence.
This intelligence has so far not been able to use insight into game A for game B. That’s the difference between Narrow AI and General AI. People who call that AI is not intelligent so far are hammering on the fact that we do not yet have a General AI.Because Narrow AI is everywhere we sit, from your washing machine to Google translate. And we are busy constructing General AI. That will take a while, even though this development is progressing exponentially. Where we are in the curve is not to say until the figurative General AI explosion comes.
My tip: Look again at this question about ten years.
One of the most heard misconceptions is that AI can develop “consciousness”.Consciousness is still not well understood, but we do know that only biological entities have consciousness. AI can simulate consciousness, but that is not the same as consciousness.
Most misconceptions about AI are easy to help out of the world by knowing its origins.AI is a self-learning version of cybernetics, a technique for self-regulating systems. The most famous example of a cybernetic system is the autopilot in airplanes.
Much misunderstanding could have been prevented if AI had been given the name describing the principle: self-learning cybernetics.The article below gives a brief explanation: