Will we ever be able to come up with a machine/device which enables us to learn....

[OinkBoink]

without (us) exerting effort?

Remember how Keanu Reeves (Neo) in The Matrix is hooked up to a computer and he learns various martial arts among other things that way (i.e. the computer teaches him)? Now, the way it is presented in The Matrix is probably too simplistic, but will we ever be able to come up with a device which allows us to memorise/memorize information, learn new skills (perhaps math(s), science etc.) etc. without us doing all the work?

 

[Spungo]

Adderall.

 

[Agent11]

I think the IV solution in THX 1138 would be a little more plausible.

'SEN 5241: You know, when I was at school, it was all very different. We used to stay in bed all the time. Combined primary economics... Combined primary economics was a bottle about this big.
[holds hands two feet apart]
SEN 5241: Took a week.
Child: Wow!'

 

[sandorski]

It might be feasible, but I ain't no Rocket Brain Engineer.

 

[John Connor]

Play learning CDs at night while you sleep everyday for weeks.

One day it may be feasible as we learn more about the brain. One day I hope that they can show your dreams on a monitor.

 

[yhelothar]

I'm a grad student working in a lab that's creating a memory prosthesis over at USC.
http://www.technologyreview.com/featuredstory/513681/memory-implants/

So far, we're able to teach a rat which of two buttons it's pushed before.
We do this by recording from the rat's brain when it learns which buttons it's pushed before. With electrodes, we can stimulate that area of the brain with the same pattern of activation, and alas, the mice "knows" what button it's pushed before.

This is still a far cry from learning math or science though. Our hippocampus that allows us to turn our memories into long term ones contains on the order of 10million neurons. We're so far able to insert 16 electrodes into the hippocampus, which so far allows for very crude learning tasks.

Also what's done so far is a simple input-output transformation. Essentially, it's treating the brain as a black box, so we are essentially mimicking its behavior without have any idea what it's doing.

There will be a long way to go in order to fully replicate all the functions of the hippocampus. Also currently the input/output transformation isn't generalizable to other tasks. That will be another hurdle to overcome. This transformation, you could imagine, also changes over time, since neurons isn't a static computational architecture, but rather a plastic one. So that will be another challenge to overcome.

I'm amazed by the ingenious people in our lab that tackles those questions though.

 

[OinkBoink]

I'm amazed by the ingenious people in our lab that tackles those questions though.

People are instinctively "amazed" by things which don't fit the norm. But being a neuroscience student, you shouldn't be (well, at least not that much). "Genius" can probably be completely explained by genetics, brain biology, the environment etc. Right?

 

[yhelothar]

The "I KNOW KUNG FU" part may be a more plausible target in the meantime though. Martial arts learned through a type of memory known as procedural memory, or colloquially, muscle memory.

This uses an entirely different system in the brain than the one you use to learn science and math, which uses the hippocampus.

The cerebellum has a very simple and organized structure and we have a good idea how it works in terms of computational machine learning algorithms. There was also a cerebellar prosthesis made that has successfully allowed rats to perform procedural memory. It looks like a very early study in this realm though, and as far as this being able to mimic much more complex forms of procedural memory, I have yet to see it. http://www.newscientist.com/article/mg21128315.700-rat-cyborg-gets-digital-cerebellum.html