"Parker said we want devices to handle unexpected circumstances"
That's the kicker. Will it be a difference engine or will it be able to make decisions when given un-thought-of situations?
Artificial brains have long been a central theme in science fiction but they inched one step closer to reality with some recent breakthroughs at the University of Southern California.
Alice Parker, a USC electrical engineering professor who has spent the last five years studying the use of carbon nanotubes to build an artificial brain, has made a recent advancement by showing that nanotechnology can in some ways behave like a neuron.
“The breakthrough is that we were able to make a transistor circuit, with the transistor formed by some carbon nanotubes, behave much like the synapse of a neuron, the bridge between the presynaptic neuron and the post synaptic neuron,” Parker told SmartPlanet.
Using a nanotube transistor created in a partner’s lab, Parker made the device behave like a synapse. “We realized if you apply an action potential to it, it creates a post-synaptic potential just like synapses in the brain would react to a spike,” Parker said.
Essentially, Parker and her team replicated a core function of the brain on a very small scale. A single neuron has 10,000 of these synapses, all acting as inputs to every neuron and that the entire brain has a hundred billion neurons - so Parker’s work managed to recreate a very small piece of a neuron, and thus a small fraction of overall brain activity.
“But what we built was important because it showed that biomimetic circuits are possible with carbon nanotube transistors. Carbon nanotube transistors have not been used in these circuits before. Building complex circuits, like synapses that can change as learning occurs, would be the next step, followed by simple neurons,” Parker said.
Assuming humans are eventually capable of building a synthetic brain, Parker explained that a key attribute of a working artificial brain is its ability to demonstrate learning and intelligence. The devices should be able to adapt over time and be smart enough to handle unexpected circumstances.
In the future, the applications of a synthetic brain include facial recognitions, vision for autonomous vehicles, robotic rescue missions and treatment for brain trauma.
Parker said sometimes it feels like doing test tube experiments. Just because she made one little thing happen, doesn’t mean she can now build an intelligent system.
So why build a synthetic brain? After hearing Ray Kurzweil talk about the possibilities of building an artificial brain, Parker admits, it motivated her to use nanotechnology to mimic the core functions of a brain. But her scientific quest began much earlier, when she read her mother’s college textbook on psychology from front to back. Plus, having a father who happened to be a co-inventor of the first synthetic vitamin B1, played a role in influencing her interest in building artificial systems.
Parker still has a ways to go before her discovery turns into a functioning artificial brain. There are certainly technical hurdles to address: Large processors require a lot of energy. With nanotechnology, the devices will shrink, become more efficient and require less energy - and might even self-assemble.
There are also ethical challenges to address. “If you have something that’s an intelligent device, does it have the rights of a human? How should it be treated. There is some discussion about the science fiction angle… about if the brains gain more intelligence than us,” Parker said, “but that is not likely.” However, she thinks, the question of how it should be treated is more important.
Photo: Tulane
Updated
Related on SmartPlanet:
