On: Understanding Deep Learning Through Neuron Deletion

[signa11]

there is a rather oldish paper called 'optimal brain damage' which also explores something _similar_. iirc, it tries to get an optimal size of the network by removing unimportant weights from it.

[Dawny33 → signa11]

Link to the same: http://yann.lecun.com/exdb/publis/pdf/lecun-90b.pdf

Also wow! Yann LeCunn has authored it

[signa11 → Dawny33]

> Link to the same

ah ! i was kind of hard pressed for time when i posted that, with the ulterior motive that some kind soul would post the link :)

thank you kindly !

[mattkrause → signa11]

Nihil sub sole novum