Artificial intelligence(AI) is actively making progress in its quest to recreate some of the models of the human brain. More specifically, AI’s focus has been centered on recreating patterns of the neocortex region of the brain which is responsible for many of the human intelligence capabilities.
As AI makes progress to recreate the neocortex, an idea that is keeping researches busy is the possibility of creating AI models that resemble transcendent abilities of the brain such as fear, creativity, imagination, aptitude among other behaviors that are key to the human identity. Many of these behaviors are the result of well-known patterns in the neocortex so is not crazy to think that AI systems can be successful imitating them. As matter of fact, there have been some AI research efforts that have shown some progress recreating transcendent abilities of the human brain. Today, I would like to focus this post in two transcendent abilities of the brain that were thought impossible to recreate by AI systems: creativity and imagination.
Human Creativity & Imagination
The neocortex specializes in recognizing patterns using a complex net of hundreds of billions o interconnected neurons. Groups of neurons in the neocortex specialized on recognizing specific patterns that could be as simple as a letter or as complex as the solution to an algebra problem. This notion of pattern recognition is the common denominator to many of the brain’s transcendent abilities.
Creativity can be described as the ability of finding great metaphor. In the context o the neocortex, a metaphor is a patterns that could be related to another pattern despite differences on details and context. The neocortex is a great metaphor association machine with the ability of recognizing over 30 billion metaphors per second.
If AI systems can be trained to be creative, they could get better at skills like creating narratives or analyzing complex problems by drawing analogies from other well-known problems.
Imagination can be described as the ability to anticipate patterns that could be related to other patterns in the future. Strategic thinking is a very clear sign of a strong imagination.
AI systems have already shown tremendous progress imitating human imagination. During game 3 of the recent match between DeepMind’s AlphaGo and Go’s number one world player Lee Sedol, AlphaGo shocked experts by finding an incredibly imaginative solution move 37. As a matter of fact, many experts believe that we will remember that specific event [game 3, move 37] as the moment when AI systems became truly creative and imaginative.
Creative and Imaginative AI
Creativity and imagination are tow of the human brain’s transcendent abilities that are possible to recreate with today’s AI technologies.
The key to enable AI creativity is to model a specific pattern [image, text, speech, etc] as a series of common constructs so that they can be used to determine its association with another patters [metaphor]. Easier said than done but an idea nonetheless. A creative AI system could be a good solution for problems such as finding treatments for a specific condition based on know treatment for other diseases.
The key to enable AI imagination is to constantly evaluate future patterns [outcomes] based on existing sets of patterns. DeepMind solution to reduce cost on Google’s data centers is a good example of AI imagination even though, of course, we will always have move 37.