Affiliation:
1. IBM Madrid Scientitic Center, Paseo de la Castellana, 4, 28046 Madrid (SPAIN)
Abstract
Neural networks are fairly straightforward to program in a matrix oriented language such as APL. The only general improvement that would benefit them would be the implementation of sparse matrics. Small networks can be trained quite easily using the standard procedures (back propagation, etc).
Publisher
Association for Computing Machinery (ACM)
Cited by
3 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献
1. U-Net CNN in APL: Exploring Zero-Framework, Zero-Library Machine Learning;Proceedings of the 9th ACM SIGPLAN International Workshop on Libraries, Languages and Compilers for Array Programming;2023-06-06
2. Adaptive learning networks in APL2;ACM SIGAPL APL Quote Quad;1993-09
3. A bibliography of APL articles on modeling and KBES;ACM SIGAPL APL Quote Quad;1990-09