arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1611.01232stat.MLcs.LG
收藏

深度信息传播

Deep Information Propagation

Google Brain(谷歌大脑);Stanford University(斯坦福大学)

Samuel S. Schoenholz, Justin Gilmer, Surya Ganguli, Jascha Sohl-Dickstein

摘要:

我们研究了未经训练的神经网络的行为,其权重和偏差是随机分布的平均场理论。我们表明存在的深度尺度,自然限制通过这些随机网络的信号传播的最大深度。我们的主要实践结果表明,随机网络可以精确地训练时,信息可以通过他们。因此,我们确定的深度尺度提供了针对特定超参数选择的网络训练深度的界限。作为一个推论,我们认为,在混沌的边缘网络,这些深度尺度之一发散。因此,任意深度的网络只能在足够接近临界的情况下进行训练。我们发现,dropout的存在破坏了有序到混沌的临界点,因此强烈限制了随机网络的最大可训练深度。最后,我们开发了一个平均场理论的反向传播,我们表明,有序和混沌相对应的区域消失和爆炸梯度分别。

英文摘要:

We study the behavior of untrained neural networks whose weights and biases are randomly distributed using mean field theory. We show the existence of depth scales that naturally limit the maximum depth of signal propagation through these random networks. Our main practical result is to show that random networks may be trained precisely when information can travel through them. Thus, the depth scales that we identify provide bounds on how deep a network may be trained for a specific choice of hyperparameters. As a corollary to this, we argue that in networks at the edge of chaos, one of these depth scales diverges. Thus arbitrarily deep networks may be trained only sufficiently close to criticality. We show that the presence of dropout destroys the order-to-chaos critical point and therefore strongly limits the maximum trainable depth for random networks. Finally, we develop a mean field theory for backpropagation and we show that the ordered and chaotic phases correspond to regions of vanishing and exploding gradient respectively.

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。

↑