List of Competition
If you can actually invent a robot with the intelligence of a human being in five years, the
design would be worth a trillion dollars. It is all a matter of how you present your ideas as to
how much of the trillion dollars you can realize.
When autonomous robots which can work on factory lines are being sold for US$50,000.00,
the market for such robots will become so large as to dwarf every other industry including oil,
utilities, airlines, automobiles and fast food. Manufacturing in China will become
"yesterday's idea" and most manufacturing will become local to the consumer.
Confabulation Theory by Robert Hecht-Nielsen seems to have the same theory of modules.
Perhaps he needs Minsky's critics. His web site is: r.ucsd.edu/index.htm
http://www.scholarpedia.org/article/Confabulation_Theory
Video at: http://CognitiveComputing2007.berkeley.edu
Hobby Robots in general seem to be progressing quite rapidly in functions and features.
Honda seems to have a good handle on the problem if you hear Edgar Koerner, President of
Honda Research Institute Europe GmbH. Here is one of their statements:
"The brain is a heterogeneously structured system, composed of numerous areas, nuclei,
types of neurons and their connections. During recent years, research in neuroscience has
provided a vast amount of new data about the architecture and mechanisms which underlie
the brain at the respective levels of analysis. In order to understand the basic principles of
information processing in the brain, this data is used to formulate integrated models of brain
functions. These models can be evaluated by transcribing them into a form which can be
simulated on a computer."
Evolved Machines
http://www.evolvedmachines.com/
These machines recognitize smells very well.
Working on visual recognition.
Claim that using actual neural network hardware speeds things up by a factor of 100.
(Cognitive Computing 2007 presentation by Paul Rhodes, PhD.)
No Free Lunch Theorem
This theorem seems to reinforce Minsky's theory that multiple systems exist for solving
different problems. There is a web site for this theorem: http://www.no-free-lunch.org/
No single algorithm is optimum for all problems.
Not sure if it is really a competitor:
Posit Sciences
www.positscience.com/
Next
If you can actually invent a robot with the intelligence of a human being in five years, the
design would be worth a trillion dollars. It is all a matter of how you present your ideas as to
how much of the trillion dollars you can realize.
When autonomous robots which can work on factory lines are being sold for US$50,000.00,
the market for such robots will become so large as to dwarf every other industry including oil,
utilities, airlines, automobiles and fast food. Manufacturing in China will become
"yesterday's idea" and most manufacturing will become local to the consumer.
Confabulation Theory by Robert Hecht-Nielsen seems to have the same theory of modules.
Perhaps he needs Minsky's critics. His web site is: r.ucsd.edu/index.htm
http://www.scholarpedia.org/article/Confabulation_Theory
Video at: http://CognitiveComputing2007.berkeley.edu
Hobby Robots in general seem to be progressing quite rapidly in functions and features.
Honda seems to have a good handle on the problem if you hear Edgar Koerner, President of
Honda Research Institute Europe GmbH. Here is one of their statements:
"The brain is a heterogeneously structured system, composed of numerous areas, nuclei,
types of neurons and their connections. During recent years, research in neuroscience has
provided a vast amount of new data about the architecture and mechanisms which underlie
the brain at the respective levels of analysis. In order to understand the basic principles of
information processing in the brain, this data is used to formulate integrated models of brain
functions. These models can be evaluated by transcribing them into a form which can be
simulated on a computer."
Evolved Machines
http://www.evolvedmachines.com/
These machines recognitize smells very well.
Working on visual recognition.
Claim that using actual neural network hardware speeds things up by a factor of 100.
(Cognitive Computing 2007 presentation by Paul Rhodes, PhD.)
No Free Lunch Theorem
This theorem seems to reinforce Minsky's theory that multiple systems exist for solving
different problems. There is a web site for this theorem: http://www.no-free-lunch.org/
No single algorithm is optimum for all problems.
Not sure if it is really a competitor:
Posit Sciences
www.positscience.com/
Next
 | |  | |  | |  |