Variants of DropConnect in Learning vector quantization networks for evaluation of classification stability
Sign Up to like & getrecommendations! Published in 2020 at "Neurocomputing"
DOI: 10.1016/j.neucom.2019.12.131
Abstract: Abstract Dropout and DropConnect are useful methods to prevent multilayer neural networks from overfitting. In addition, it turns out that these tools can also be used to estimate the stability of networks regarding disturbances. Prototype… read more here.
Keywords: vector; classification; variants dropconnect; non linear ... See more keywords