Slashdot Mirror
Can We Build a Human Brain Into a Microchip?
destinyland writes "Can we imprint the circuitry of the human brain onto a silicon chip? It requires a computational capacity of 36.8 petaflops — a thousand trillion floating point operations per second — but a team of European scientists has already simulated 200,000 neurons linked up by 50 million synaptic connections. And their brain-chip is scaleable, with plans to create a superchip mimicking 1 billion neurons and 10 trillion synapses. Unfortunately, the human brain has 22 billion neurons and 220 trillion synapses. Just remember Ray Kurzweil's argument: once a machine can achieve a human level of intelligence — it can also exceed it."
Just remember Ray Kurzweil's argument: once a machine can achieve a human level of intelligence â" it can also exceed it.
Ray Kurzweil is a brilliant computer scientist and brought us many improvements -- maybe even the invention of -- the electronic musical keyboard.
But that is not his argument. I laughed when I read that as the concept was presented to me in sci-fi novels before Kurzweil's time. The earliest I (or Wikipedia) can trace the intelligence explosion theory back to is Irving John Good who, in 1965, said:
Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines; there would then unquestionably be an 'intelligence explosion,' and the intelligence of man would be left far behind. Thus the first ultraintelligent machine is the last invention that man need ever make.
This was popularized by Vernor Vinge which is where I recalled reading about it. There are many reasons to celebrate Raymond Kurzweil. In my opinion, his is "work" in nutrition and his near-religion called futurology are not in those reasons. He has become a vocal proponent of a dream to become god-like. I do not share that dream and I wish him the best of luck in his endeavors. I just cringe every time I read of the "singularity being near" or the ability to live forever coming about. If it's going to happen, just sit back and let it happen. I feel he has done a great disservice to the field of artificial intelligence by promising unrealistic things in interviews to the lay person. Disappointment is a sure fire way to get yourself branded as a snake oil salesman religious nut.
Predictions for the future are for sci-fi books and movies, don't get into the habit of being a scientist in an interview with a reputable magazine or web site telling them what is about to happen. Example:
Kurzweil projects that between now and 2050 technology will become so advanced that medical advances will allow people to radically extend their lifespans while preserving and even improving quality of life as they age. The aging process could at first be slowed, then halted, and then reversed as newer and better medical technologies became available. Kurzweil argues that much of this will be a fruit of advances in medical nanotechnology, which will allow microscopic machines to travel through one's body and repair all types of damage at the cellular level.
And that's easily criticized:
Biologist P.Z. Myers has criticized Kurzweil's predictions as being based on "New Age spiritualism" rather than science and says that Kurzweil does not understand basic biology. Myers also claims that Kurzweil picks and chooses events that appear to demonstrate his claim of exponential technological increase leading up to a singularity, and ignores events that do not.
My work here is dung.
Something like this will be possible one day, but my layperson's understanding of how the brain works is fundamentally different from how computers work. The hard-wired CPU/RAM model is just not a perfect parallel, so while we can and will improve on machines that learn, it's going to be different from the wetware that is constantly growing, changing, forming new connections and interacting with internal, external and imagined stimuli.
The CB App. What's your 20?
It's the reconfigurable nature of the human brain that's unique and powerful. If all you did was take one person, listed all of the skills of that person -- all of the things he knew; all of the skills in smell, touch, sight and taste; all of the cognitive reasoning ability -- then you could create a chip to simulate those skills. Algorithms for image recognition, feature extraction, speech recognition, etc. are all available that are very very close to what humans can do.
But the thing that separates humans is that it didn't take hundreds of years of mathematical development to come up with these algorithms. The human brain develops these algorithms through changes in its structure from birth. At about age 10, speech recognition specialized and tailored to the dialect, language and tones that the person hears has developed on its own.
That type of structural formation and learning is what would need to happen in silicon to make a truly intelligent machine. Neuron clusters emulated using transistors would need to be able to dynamically form connections to other neuron clusters. There'd have to be some type of distributed learning algorithm encoded in the operation of each individual neuron.
Speech recognition is easy. Image recognition is easy. Developing a distributed, scalable, self-modifying architecture that can learn all of those and more on its own with nothing more than training samples is the difficult part.
As one of my professors once said: "How do we go from billions of neural synapses to midget wrestling?" While amusing, it points out one of our great unknowns. Biologists and neuroscientists (some psychologists) understand things at the synapse level, and how the chained firing happens in neurons. Then psychologists understand normal behavior by examining abnormal behavior, but that's at a much higher level. We simply don't know how to map out what's in between.
The Ship of Theseus, also known as Theseus's paradox, is a paradox that raises the question of whether an object which has had all its component parts replaced remains fundamentally the same object.
BTW, current estimates are more like 100 billion neurons and upwards of 300-500 trillion synaptic connections.
However, numbers aside, the human brain is not merely a complex collection of neurons and interconnected synapses. Complexity is only one very basic factor, another, more critical, factor is organization. We don't even know where to start in the organization of these artificial neural networks to emulate a human brain.
WARNING! COMPUTER ANALOGY: It's not the number and density of interconnected transistors that make a Xeon, it's the organization.