arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1506.08700stat.MLcs.LG
收藏

脱落作为数据扩充

Dropout as data augmentation

Université de Montréal(蒙特利尔大学);Goethe University Frankfurt(法兰克福歌德大学);CIFAR(加拿大高等研究院)

Xavier Bouthillier, Kishore Konda, Pascal Vincent, Roland Memisevic

摘要:

Dropout通常被解释为将大量共享参数的模型打包。我们表明,在网络中使用dropout也可以被解释为一种没有领域知识的输入空间中的数据增强。我们提出了一种方法,将网络中的丢弃噪声投射回输入空间,从而生成训练数据的增强版本,并且我们证明了在增强样本上训练确定性网络会产生类似的结果。最后,我们提出了一个新的dropout噪声方案的基础上,我们的观察,并表明它改善了辍学的结果,而不增加显着的计算成本。

英文摘要:

Dropout is typically interpreted as bagging a large number of models sharing parameters. We show that using dropout in a network can also be interpreted as a kind of data augmentation in the input space without domain knowledge. We present an approach to projecting the dropout noise within a network back into the input space, thereby generating augmented versions of the training data, and we show that training a deterministic network on the augmented samples yields similar results. Finally, we propose a new dropout noise scheme based on our observations and show that it improves dropout results without adding significant computational cost.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑