Page Title
 |
| |
 |
| |
 |
| |
 |
On p. 69 of The Singularity is Near, Kurzweil printed the Hans Moravec chart showing the
power of $1000 worth of computing. The 1995 trendline predicts that a $1000 computer will
have the power of a mouse brain by 2007. As of 2007, it takes a multimillion dollar
supercomputer to equal one half of a mouse brain operating at 1/10th speed. See IBM
Almaden: http://www.almaden.ibm.com/cs/people/dmodha/
In another report, "The scientists ran a "cortical simulator" that was as
big and as complex as half of a mouse brain on the BlueGene L
supercomputer.
In other smaller simulations the researchers say they have seen
characteristics of thought patterns observed in real mouse brains.
Now the team is tuning the simulation to make it run faster and to make it
more like a real mouse brain. " http://news.bbc.co.uk/2/hi/technology/6600965.stm
"Half a real mouse brain is thought to have about eight million neurons
each one of which can have up to 8,000 synapses, or connections, with
other nerve fibres.
Modelling such a system, the trio wrote, puts "tremendous constraints on
computation, communication and memory capacity of any computing platform".
The team, from the IBM Almaden Research Lab and the University of Nevada,
ran the simulation on a BlueGene L supercomputer that had 4,096
processors, each one of which used 256MB of memory.
Using this machine the researchers created half a virtual mouse brain that
had 8,000,000 neurons that had up to 6,300 synapses.
The vast complexity of the simulation meant that it was only run for 10
seconds at a speed ten times slower than real life - the equivalent of one
second in a real mouse brain."
It doesn't look like we are even close to what Moravec and Kurzweil predicted.
One could argue back that a modern laptop has the computational ability of a mouse. But it
cannot recognize cheese, nor can it navigate an alley full of trash cans like a mouse can do.
Michio Kaku says that AI is 100 years off because Moore's Law is about 20 years from
collapsing:
http://www.youtube.com/watch?v=PW8rgKLPHMg
But, Dr. Kaku, the human brain stuffs intelligence in three pounds.
This video is about 100 times as popular as the Memristor Symposium.
Greg Snyder spoke at The Memristor Symposium and argues that the feed-forward neural
networks of the 1980s were valuable, but limited. They are still used to recognize postal zip
codes and read numbers off of checks, but widespread use for other things has not yet
occurred. Apparently what is needed is more advanced neural networks. At the same time,
using traditional von Neumann architecture has led us to simulating a mouse brain using 40
kilowatts of supercomputer. Meanwhile, advances in memristor design indicate that neural
networks can be created in a compact manner which can also be stacked in the third
dimension and which allow devices which are not being switched to not have to consume
power. At the time of the symposium, memristors were limited to 10^7 cycles before wearing
out. Perhaps meminductors will be invented that last longer due to magnetism not "wearing
out" as quickly as the actual chemical transport used in titanium oxide memristors at this time.
Next
power of $1000 worth of computing. The 1995 trendline predicts that a $1000 computer will
have the power of a mouse brain by 2007. As of 2007, it takes a multimillion dollar
supercomputer to equal one half of a mouse brain operating at 1/10th speed. See IBM
Almaden: http://www.almaden.ibm.com/cs/people/dmodha/
In another report, "The scientists ran a "cortical simulator" that was as
big and as complex as half of a mouse brain on the BlueGene L
supercomputer.
In other smaller simulations the researchers say they have seen
characteristics of thought patterns observed in real mouse brains.
Now the team is tuning the simulation to make it run faster and to make it
more like a real mouse brain. " http://news.bbc.co.uk/2/hi/technology/6600965.stm
"Half a real mouse brain is thought to have about eight million neurons
each one of which can have up to 8,000 synapses, or connections, with
other nerve fibres.
Modelling such a system, the trio wrote, puts "tremendous constraints on
computation, communication and memory capacity of any computing platform".
The team, from the IBM Almaden Research Lab and the University of Nevada,
ran the simulation on a BlueGene L supercomputer that had 4,096
processors, each one of which used 256MB of memory.
Using this machine the researchers created half a virtual mouse brain that
had 8,000,000 neurons that had up to 6,300 synapses.
The vast complexity of the simulation meant that it was only run for 10
seconds at a speed ten times slower than real life - the equivalent of one
second in a real mouse brain."
It doesn't look like we are even close to what Moravec and Kurzweil predicted.
One could argue back that a modern laptop has the computational ability of a mouse. But it
cannot recognize cheese, nor can it navigate an alley full of trash cans like a mouse can do.
Michio Kaku says that AI is 100 years off because Moore's Law is about 20 years from
collapsing:
http://www.youtube.com/watch?v=PW8rgKLPHMg
But, Dr. Kaku, the human brain stuffs intelligence in three pounds.
This video is about 100 times as popular as the Memristor Symposium.
Greg Snyder spoke at The Memristor Symposium and argues that the feed-forward neural
networks of the 1980s were valuable, but limited. They are still used to recognize postal zip
codes and read numbers off of checks, but widespread use for other things has not yet
occurred. Apparently what is needed is more advanced neural networks. At the same time,
using traditional von Neumann architecture has led us to simulating a mouse brain using 40
kilowatts of supercomputer. Meanwhile, advances in memristor design indicate that neural
networks can be created in a compact manner which can also be stacked in the third
dimension and which allow devices which are not being switched to not have to consume
power. At the time of the symposium, memristors were limited to 10^7 cycles before wearing
out. Perhaps meminductors will be invented that last longer due to magnetism not "wearing
out" as quickly as the actual chemical transport used in titanium oxide memristors at this time.
Next