New answers tagged math
0
votes
Accepted
Is there a mathematical proof that a binary neural network can approximate any function with arbitrary accuracy?
Yes, for a broad class, they actually do, with probability 1:
[i] Wang, Yanzhi, et al. "Universal approximation property and equivalence of stochastic computing-based neural networks and binary ...
- 156
Top 50 recent answers are included
Related Tags
math × 235machine-learning × 58
neural-networks × 45
deep-learning × 39
reinforcement-learning × 34
papers × 15
backpropagation × 14
terminology × 13
objective-functions × 13
notation × 13
gradient-descent × 12
proofs × 12
convolutional-neural-networks × 10
definitions × 10
activation-functions × 10
generative-adversarial-networks × 8
optimization × 8
function-approximation × 8
derivative × 8
natural-language-processing × 7
reference-request × 7
datasets × 7
applications × 7
policy-gradients × 7
gradient × 7