| [ Java | No Java | Text Only ] |
| Thoughts on Thought |
| Language |
| Brain & Computer |
| Pattern-Recognition |
| AI Components |
| Applications |
| To Be Human |
| The Future |
| FAQs |
The Future--Faster Computing |
[ Traditional Computing | Parallel Processing ] Traditional Computing
Another way to make a computer faster is by using light instead of electricity to transmit information through a circuit. Since light travels faster than electricity(in fact, the speed of the light is the fastest thing in the universe), an optical computer would be able to process the large data produced for the system's AI much faster. Experiments with photonic crystals are currently underway that allow scientists guide where light travels. Since light would not need a conduit to direct its path, there would be little resistance and heat to worry about. Plus, because light can be composed of many different wavelengths, a large amount of data can be transmitted in the same beam of light! Still a third solution to speedier computing are bio-chips--computer chips that have biological components integrated into its circuits. Early prototypes of bio-chips have been produced in various research institutions like the California Institute of Technology. In this scheme, bio-chips are not built but are grown and the principles governing biological neurons would somewhat apply to these special devices. These chips may also bring AI researchers one step closer in constructing neural networks that function like the human brain. [ Top ] Parallel ProcessingA faster alternative to building faster and more complex chips is to produce microprocessors that work together simultaneously. Currently, most computers work the same way as the first generation of computers did--information is processed by the computer one binary bit at a time called serial computing. The concept of parallel processing is that the entire data is broken up into small and manageable pieces which are distributed to many simple computers that digest the information and return a result at the same time. Together, the many processors would be able to calculate exponentially faster than the traditional serial computer. The shrinking of computers may go beyond the molecular level and into the world of electrons. Quantum computers may be the answer to computing speed by combining the benefits of miniaturization with parallel processing. The electrons will store and calculate the binary bits of data from their direction of what quantum physicists call "spin." It can spin one way to hold a "0" and the other way to store "1". However, since electrons spin both directions at the same time, one electron, called a qubit(short for "quantum bit"), can theoretically hold both "0" and "1" at the same time. Existing in a medium similar to coffee and controlled by magnetic beams, two electrons would hold for bits, three would hold sixth and so forth. Thus, each electron performs its own calculation and returns a value in parallel. Presently, the interpretation of the electrons without changing their spins is a challenge that is being worked on today. |
he present day step
towards AI is making faster computers that can handle more data in less time. The
easiest way to achieve this is by making computer circuits smaller and smaller so that it
takes less time for the electrical impulses to travel from one region to another.
Perhaps, computers will one day become so small that they are as no bigger than a few
molecules in which the age of molecular computing will begin. However, with less
material for the heat generated by the movement of electricity through the metals to
dissipate, computers would literally melt. Research in sup