Statistical-Query Lower Bounds via Functional Gradients

NIPS 2020, 2020.

Cited by: 3|Views13
EI
Weibo:
(ICML 2020) on the SQ dimension of functions computed by two-layer neural networks

Abstract:

We give the first statistical-query lower bounds for agnostically learning any non-polynomial activation with respect to Gaussian marginals (e.g., ReLU, sigmoid, sign). For the specific problem of ReLU regression (equivalently, agnostically learning a ReLU), we show that any statistical-query algorithm with tolerance $n^{-\Theta(\epsilo...More

Code:

Data:

Your rating :
0

 

Tags
Comments