Graphene Flakes Facilitate Neuromorphic Chips (ieee.org)

An anonymous reader writes: One of the hot areas of semiconductor research right now is the creation of so-called neuromorphic chips — processors whose transistors are networked in such a way to imitate how neurons interact. "One way of building such transistors is to construct them of lasers that rely on an encoding approach called "spiking." Depending on the input, the laser will either provide a brief spike in its output of photons or not respond at all. Instead of using the on or off state of the transistor to represent the 1s and 0s of digital data, these neural transistors rely on the time intervals between spikes." Now, research published in Nature Scientific Reports has shown how to stabilize these laser spikes, so that they're responsive at picosecond intervals. "The team achieved this by placing a tiny piece of graphene inside a semiconductor laser. The graphene acts as a 'saturable absorber,' soaking up photons and then emitting them in a quick burst. Graphene, it turns out, makes a good saturable absorber because it can take up and release a lot of photons extremely fast, and it works at any wavelength; so lasers emitting different colors could be used simultaneously, without interfering with each other—speeding processing."

Comments

I need this

[PopeRatzo]

If neuromorphic chips can help my PC run Just Cause 3 better, I'm all for them.

Re:I need this

[Fwipp]

With neuromorphic graphene flakes, we can accurately model and simulate the mental states of dozen of AI people at a time, complete with artificial memories, fears, and dreams. And, if you purchase the dedicated LaZer X coprocessor card, not only will we simulate the internal brain states of your foes, but also that of their loved ones - blowing people up has never been so real!

Re:I need this

[cheater512]

Put two LaZer X cards in your computer and the second one will simulate your own brain's state after a brief scan, and will play the game just as you would but without having to actually be near the PC.

Useless

[Anonymous Coward]

I'm not interested until it's 3D printed.

Star Trek

[peragrin]

I was just watching Star trek and then I read slashdot. I am not sure which had the worse technobabble.

So when can I get my Brain Transplant?

[argee]

One of my friends told me all I need is a Brain Transplant and a Face Lift, and I'd be OK.
So, how much $$ ?

Easier summary:

[Y.A.A.P.]

It's been discovered that Graphene can be used as an optical capacitor.

Add in the rest about chips patterned after the brain but otherwise have no other connection to it.

Re:Easier summary:

[Anonymous Coward]

Ahh sweet Graphene. Is there nothing you can't do?

Re:Easier summary:

[NoNonAlphaCharsHere]

I wonder what happens when you pump 1.21 gigawatts into one.

Is there anything it cannot do?

[roman_mir]

Maybe Grapheme should run for the president?

Optical Clock to Q

[Anonymous Coward]

Optical capacitance could open the door an interesting new avenue for computational parallelism if indeed it works across a wide spectrum. Consider the gate allowing an optical capacitor to drive its output to be your system clock, you simply base your system clock frequency on the maximum latency for the logic required to determine the input state for any given band in the usable spectrum. Suddenly building an ALU which can perform a different parallel operation for each band within the usable optical spectrum becomes trivial.

time duration

[globaljustin]

I thought it was interesting that they were using time division *between* pulses instead of binary.

The brain does process on a continual basis so it makes sense. I don't know if time-division multiplexing neurons is something every neuroscientist knows about but I thought it was at least noteworthy.

Re:time duration

[110010001000]

That isn't time division multiplexing.

Re:time duration

[globaljustin]

yeah technically it's not multiplexing signals

Re:time duration

[Anonymous Coward]

It's pulse position modulation(likely amplitude modulation as well).

http://antranik.org/wp-content/uploads/2012/04/synapse.jpg
https://en.wikipedia.org/wiki/Ligand-gated_ion_channel

Khan Academy has some pretty good videos on how synaptic depression and activation contribute to hebbian learning.

Modern neuromorphic chips use a Spiking Neural Network architecture to maximize realism.

Skynet

[sycodon]

Sounds like the work of Cyberdyne Systems/

Looks Far Away

[JimSadler]

Bringing this technology into common products, sounds to me that it might take decades. It would be a nice surprise if it can be useful sooner than I suspect.

Re:Looks Far Away

[Anonymous Coward]

Use FPGAs

How long has it been?

[Anonymous Coward]

Is it ten years since graphene was advertised as the wonder material? Since then, we keep hearing that it has potential applications nothing short of terrific. Since then, all those applications seem to remain firmly in the potential stage.

Star Wars

[O-Deka-K]

The first thing that came to mind: "I don't know where you get your delusions, _____ _____." (fill in the blanks)