Neurocomputing: Picking the human brain
As an alternative form of information processing, neurocomputing is fast becoming an established discipline, and some neural networks are already on the market. Neural networks are good at some things that conventional computers are bad at. They do well, for instance, at solving complex pattern-recognition problems implicit in understanding continuous speech, identifying handwritten characters, and determining that a target seen from different angles is in fact one and the same object. Neural networks parallel-process immense quantities of information. Yet for a long time the only way to implement them was by simulating them laboriously, inefficiently, and at great expense on standard, serial computers. That situation is changing. Neurocomputers - hardware on which neural networks can be implemented efficiently - have reached the prototype stage at several companies, and some are already commercially available. All are coprocessor boards that plug into conventional machines. Developers include Hecht-Nielsen Neurocomputer Corp. (HNC), IBM Corp., Science Applications International Corp. (SAIC), Texas Instruments Corp., and TRW Inc. Meanwhile, researchers at Boston University, the California Institute of Technology, the Helsinki University of Technology, Johns Hopkins University, the University of California at San Diego, and other universities have been investigating the theory behind neural networks and exploring their potential to solve problems that have stumped algorithmic computing for decades. This article discusses neurocomputing.
- Research Organization:
- Hecht-Nielsen Neurocomputer Corp. (US)
- OSTI ID:
- 7169707
- Journal Information:
- IEEE Spectrum; (United States), Journal Name: IEEE Spectrum; (United States) Vol. 25:3; ISSN IEESA
- Country of Publication:
- United States
- Language:
- English
Similar Records
The SAIC delta neurocomputer architecture
A simple VLSI architecture for neurocomputing