But there is another important point here. Intelligence was not the last theory of evolution, a thing to do. Instead, they evolved in a variety of ways from a few complex solutions that allowed life to survive and come into the future. Intelligence is the highest priority right now in terms of continuity and openness. In this way, evolution is a far cry from the way people view it – as a means to an end.
It is this open end, which is reflected in the complexities of POET, which Clune and others believe can bring about new forms of AI. For decades AI researchers have been trying to create algorithms to mimic human intelligence, but real success can come from building algorithms that try to mimic the end of endless problems — and sit back to see what it looks like.
Researchers are using machines to study on their own, training them to find solutions to complex problems in the field, such as making machines that can learn one task at a time or deal with challenges they have never encountered. Some now think that using this method and walking it is probably the best way to do it. “We could start an algorithm that initially doesn’t have a lot of intelligence inside, and we’ll look at bootstrap itself right up to the AGI,” says Clune.
The truth is that at the moment, AGI is still speculative. But mainly because no one knows how to make it. Advances in AI are insignificant and man-made, advances that affect tweaks to existing systems or algorithms, which give greater leaps to performance or accuracy. Clune is known for this endeavor in an attempt to find the smartest architecture to create without knowing what you are looking for or how many places you would like. And that is just the beginning. “Sometimes, we have to work on Herculean to get them together,” he says.
Asking AI to find and make our own constructions and paradigm changes. I say we want to make smart machines, but we don’t care what it looks like – just give us everything that will work.
Although the AGI is not implemented, the self-training approach can change the type of AI generated. The world needs more than the best Go player, says Clune. For him, developing high-tech machines means creating machines that create their own problems, solve them, and create new ones. A POET makes this point. Clune thinks of a machine that teaches a bot to walk, and then to play hopscotch, then maybe to play Go. “Then he probably learns math and starts to cause his problems,” she says. “The machines continue to make new ones, and the sky is the limit where it can go.”
Source link
