All translations
Enter a message name below to show all available translations.
Found 3 translations.
| Name | Current message text |
|---|---|
| h English (en) | '''{{Term|spatial dropout}}''' (Tompson et al., 2015) is particularly important for {{Term|convolutional neural network|convolutional networks}}. Standard dropout on convolutional feature {{Term|map|maps}} is ineffective because adjacent {{Term|activation function|activations}} are highly correlated; dropping individual pixels still leaves redundant spatial information. {{Term|spatial dropout}} instead drops entire {{Term|feature map|channels}}, forcing the network to use diverse feature representations. |
| h Spanish (es) | El '''{{Term|spatial dropout|dropout espacial}}''' (Tompson et al., 2015) es particularmente importante para las {{Term|convolutional neural network|redes convolucionales}}. El dropout estándar en {{Term|map|mapas}} de características convolucionales es ineficaz porque las {{Term|activation function|activaciones}} adyacentes están altamente correlacionadas; descartar píxeles individuales aún deja información espacial redundante. El {{Term|spatial dropout|dropout espacial}} en cambio descarta {{Term|feature map|canales}} enteros, obligando a la red a usar representaciones de características diversas. |
| h Chinese (zh) | '''{{Term|spatial dropout|空间 dropout}}'''(Tompson 等,2015)对于{{Term|convolutional neural network|卷积网络}}尤其重要。在卷积特征{{Term|map|图}}上使用标准 dropout 是无效的,因为相邻的{{Term|activation function|激活值}}高度相关;丢弃单个像素仍然会留下冗余的空间信息。{{Term|spatial dropout|空间 dropout}}转而丢弃整个{{Term|feature map|通道}},迫使网络使用多样化的特征表示。 |