Architecture Studies and System Demonstrations of Optical Parallel Processor for AI (Artificial Intelligence) and NI (Neural Intelligence)
Abstract
During the last six months we have applied the results of our studies on existing parallel computing architectures for AI and NI to develop the Programmable OptoElectronic Multiprocessor (POEM) architecture. Our goal was design a scalable architecture suitable for AI and ultimately for NI that will take full advantage of the hybrid nature of opto-electronic technologies. In the POEM system this is achieved by implementing all communication using phontonics and all logic their local interconnections using electronics.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1988
- Accession Number
- ADA203241
Entities
People
- Sing H. Lee
Organizations
- University of California, San Diego