19-03-2011, 12:22 PM
neurocomp.ppt (Size: 58.5 KB / Downloads: 54)
Introduction
• Neurocomputers, “Computers with Brain”
• Artificial intelligence allow a computer to get really smart
• neuro-electronics
• Trents in improvements
[/b]Reasons for Doing Study on Neurocomputers
• main reason
• advantages of neural networks and neurocomputers
• repetition of new and old information
• neural networks and neurocomputers better than more conventional kinds of software and hardware
[b]Advantages of Neurocomputers
• demonstrations of adaptiveness
• learning
• associative memory
• optimization and classification of information
Breakthroughs in Neurocomputing
• Since the first working model of neurocomputing was invented only five years ago, this technology is still very much in its infancy
• Some automated procedure, whether enzymatic, mechanical, or electrical, must facilitate the placement of neurons in their wells without a human technician guiding each separate cell.
• Eve Marder of Brandeis University developed a computer program to translate neural communication
Processing Characteristics
• Self-organization
• Multiple States
Neurocomputers Inside Supercomputers
• Exists in the form of the program that implements the neural network algorithm of the optimal loading distribution between processors
• The neural network is considered as the component part of the optimization compiler
• For maximal efficiency for the distributed information processing system, one needs to equalize the load feeding to each processor
Benefits
• Medicine
• Quantum Mechanics
• Processing Efficiency
Consequences of Neurocomputing
• Direction of synthesizing intelligent life from machine parts
• The Matrix?
• The rational thinking of manipulated brain cells combined with the memory and storage capacity of traditional computers would create a mind far superior to any natural human mind.
Conclusion
• Bare–bone architecture
• -In about 10 years or so neurocomputers will go commercial.