[infurl]

https://thenextweb.com/neural/2020/09/03/the-fourth-generation-of-ai-is-here-and-its-called-artificial-intuition/

In order to have true “artificial intelligence,” we need machines that can “think” on their own, especially when faced with an unfamiliar situation. We need AI that can not just analyze the data it is shown, but express a “gut feeling” when something doesn’t add up. In short, we need AI that can mimic human intuition.

First, instead of building a quantitative model to process the data, artificial intuition applies a qualitative model. It analyzes the dataset and develops a contextual language that represents the overall configuration of what it observes. This language uses a variety of mathematical models such as  matrices, euclidean and multidimensional space, linear equations and eigenvalues to represent the “big picture.”  If you envision the big picture as a giant puzzle, artificial intuition is able to see the completed puzzle right from the start, and then work backward to fill in the gaps based on the interrelationships of the eigenvectors.

[frankinstien]

I think most might confuse intuition with unconscious processing. Here's a paper that looks into this kind of phenomena. Effectively what our brains can process doesn't necessarily require humans to be aware of. I know from personal experience that something I couldn't solve can suddenly pop up hours, days, and even weeks later.  It's almost as if some background processing is going on and when it's done it makes itself known to what we call consciousness.  There are those situations where I know that I know but I don't have the details of what that is! So not all information is made available to conscious awareness and it does appear that simply signaling that a solution is available is all we need. The example explained in the article, financial fraud, was more of a pattern detection process rather than intuition.

[yotamarker]

I can't follow what you mean by intuition. tbh ngl
perhaps you could provide examples ?

[frankinstien]

I can't follow what you mean by intuition. tbh ngl
perhaps you could provide examples ?

I'm assuming you are responding to my post? I believe that the article delved into areas where it was describing unconscious processing, whereas I look at intuition more as actionable thought that is like a reflex. As when someone first is learning a task they have to think about it all the details but once its engrained it's second nature.

[LOCKSUIT]

The article does not explain how Artificial Intuition actually works. I'm upset. It just says it sees the big picture and then fill in gaps. Seriously. I'm assuming it said also that 2 features that both share anomalies are more related probably ex. cat has alien toe on back, dog has hole in tail. That's the only thing I can take away from the article maybe.

Will check these soon.
https://www.scnsoft.com/blog/4-types-of-data-analytics
https://medium.com/@liyenz/types-of-analytics-649acafe8966

[Dee]

how Artificial Intuition actually works.

i don't get much about what 'artificial intuition' means, is it a feature of AGI, kinda a sum of multiple conclusions?

[HS]

This may sound like nonsense, I haven’t fully worked this idea out, it’s something like this. As a simple machine, if the shortest distance between two points contains an obstacle, the path between the two points has a trade off. You can plan a simple route of two straight lines at an angle to get around the obstacle, which doesn’t require much calculation, but is physically inefficient. Or you can go around in a smooth curve, which is the most physically efficient, but requires more calculation. The best solution would be efficient with both plans and actions.

Intuition makes it so that the path around an obstacle doesn’t have to be a deliberate calculation, it can be the result of a different process, like a bike being compelled to turn in a specific arc to maintain balance. You can’t really stop due to “does not compute” like robots do, or you’ll fall over. If the goal is to keep an internal progression of mental states alive, you could bypass external inconsistencies. Imagine pairing musicality with mathematical reasoning. Just as certain chords go together, certain internal states, of neurotransmitters, emotions, and reasoning, could lead into each other, and some of those progressions could harmonize. Each possible effective progression of chords would be like a personality signature, and represent a different path around the obstacle. Remembering effective patterns could let you predict some of the future by self determination.

Since you would be following along with a kind of harmonic momentum, there would be an energetic incentive/advantage to stay in the groove, as it were. The burden of calculation would be reduced in the present. The previous states/chords would help to determine current states, and these would help determine fitting options for the future. Sometimes a situation is more digital in nature, sometimes it is more analogue. The ability to operate on both layers would speed things up. Allowing you to bend your path arbitrarily around an obstacle without inventing trigonometry, and then reason with a bike lock using numbers instead of martial arts. The weakness of robots is they can’t get away from precise calculations. While the weakness of humans is not being able to go far enough into precise calculations. These traits define/confine our regions of ability. We have created tools which increase our ability to compute and remember. Similarly, the hard and angular reasoning of robots could benefit from the undefined/emergent tools existing on the other side of the ∏ fence.