The most popular and comprehensive Open Source ECM platform
Let us know more
Artificial Intelligence: Can Algorithms Have Imagination?
You frequently hear the questions: ‘Do computers think’, ‘Can AI be considered conscious?’, ‘Can an algorithm have imagination?’
It’s clear that if there’s any thinking going on by AI and computers, the mechanics of doing that are different than they are for humans. But, despite that, even the current generation of AI algorithms like OpenAI’s GPT-3 and DALL-E can create pretty impressive results.
Consider these computer-generated sketches based off of a text prompt. I think there would be general agreement that these would be considered ‘imaginative’ if they would have been created by a human.
The potential for Imaginative AI seems strong. But there is still no general formula for building an AI algorithm that can demonstrate imagination.
A group of researchers at USC is trying to develop a general approach for AI algorithms. The approach is to develop a something like a decision tree of potential attributes which act as input parameters, and from that, to apply ‘imagination’ that can envision the consequences of changing the attributes.
Yunhao Ge, PhD student at USC, said that “we were inspired by human visual generalization capabilities to try to simulate human imagination in machines. Humans can separate their learned knowledge by attributes—for instance, shape, pose, position, color—and then recombine them to imagine a new object. We attempt to simulate this process using neural networks.”
