backpropとは 意味・読み方・使い方

名詞

backprop (uncountable)

1. backpropagation
   • 2015, Sanjeev Arora, Yingyu Liang, Tengyu Ma, “Why are deep nets reversible: A simple theory, with implications for training”, in arXiv‎^[1]:

     The generative model suggests a simple modification for training---use an input to produce several synthetic inputs with the same label, and include them in the backprop training.

   • 2020, Timothy P. Lillicrap, Adam Santoro, Luke Marris, Colin J. Akerman and Geoffrey Hinton, “Backpropagation and the brain”, in Nature‎^[2]:

     In machine learning, backpropagation of error (‘backprop’) is the algorithm most often used to train deep neural networks and is the most successful learning procedure for these networks.