Simulating the Mind

Post Reply New Topic

The technicalities of that field are quite interesting..

We can probably create human level artificial intelligence within a short time (10-20 years) from which point the bridge to superhuman intelligence should be quite short. Numenta's technology is great; modeling software after the neocortex should provide computing solutions for information sorting problems computers presently find difficult, but the next step would be to integrate this into a larger architecture consisting of several similar algorithms performing other simple functions. This cognitive framework would not be limited to simple recursive algorithms like most AI programs do, but would have an immense scope of intelligence and versatility, a true General Intelligence.

I'm on the side of Ben Goertzel in proposing that the system would need to be implemented in a 3D environment and allowed to have a learning experience similar to the human one. It's interesting to note that while most AI programs of today are designed to display results within very short periods of time, in reality it takes humans years of mumbling before any form of language develops, so such an AI program should not be concerned so much with rigid speed as with the broad and subtle efficiency with which humans use to develop such fluid and highly nuanced cognition.

Anyway, I'm thinking with the advent of superintelligence, progress in this area will become very rapid, considering the effects it will have on nanotechnology and vice versa. Augmenting humans would not be a problem with sufficiently advanced nanotechnology; it's quite interesting the things they're proposing now.

I cannot word my idea's properly, but I am thinking what if a human already in existance were to become this.

I think the mistake they are making with AI is focusing too much on all aspects at once. AI would be useful but simulating the human mind including it's flaws would be pointless and less likely to succeed. Estimate is probably a better word anyway. I would be interested in focusing on trying to estimate judgment (by a logical definition), but also motor intelligence is extremely important. But not just an excuse to stuff a soft toy with animatronics. Another part of AI that I think is counter productive is the idea the processing capacity is the answer to everything. Processing capacity itself only has x efficiency. At the current rate a lot of complex problems won't be solved this way. There are various things that you could to improve efficiency. Judgment ultimately is the key to make efficient. I think a good place to start is the game Go/ Wei Chi
--> Computer Go

Question:
How would you implement timing in a neural circuit or a combination of other computing paradigms?

--> Psychomotor neurons are very active at timing things and is based on neuromuscular metabolism etc..
The mechanisms in my hand movements for an example are very much involved with spatial and temporal input and output.

One interesting other thing on my mind that I have developed recently is what if there was a model to generalized mechanisms in behaviors when the brain is functioning a certain way and to try to understand the role that the neurotransmitters have in the purpose of those behaviors.

An example is that my generalized veiw of depressive states is linked to the brains error detection system looking for the error to be resolved. The same goes with any of the other states linked to fear and anxiety.

So I think that this could explain some about stuff like OCD and aspergers autism right? But Ben Geortzel had that dynamical psychology thing that I visited before that I noticed so he might have enough crack-pot theories of the minds' internal mechanisms.

There is a therory that depression, etc is due to lost messages. It has to jump a gap to reach the receptors.