[argonaut]

The irony is that dropout is increasingly and often viewed as unnecessary in recent deep learning work.