Abstract
Neural Networks are not programmed - they are trained. While not appropriate for numeric applications, neural networks perform well in domains rife with "fuzzy knowledge", that is, knowledge not easily encapsulated.
© 1990 Optical Society of America
PDF ArticleMore Like This
A. K. Krishnamurthy, S. A. Collins, S. C. Ahalt, and D. Stewart
MJJ4 OSA Annual Meeting (FIO) 1990
Mhanna Mhanna, Mohamed Sy, and Aamir Farooq
JTh3A.49 CLEO: Applications and Technology (CLEO:A&T) 2021
John Peurifoy, Yichen Shen, Yi Yang, Li Jing, Fidel Cano-Renteria, John Joannopoulos, Max Tegmark, and Marin Soljačić
FTh4A.4 Frontiers in Optics (FiO) 2017