4 ms·
> First, this doesn't mean that a network can be used to exactly compute any function. Rather, we can get an approximation that is as good as we want. By increa
by hcs 7y ago
> First, this doesn't mean that a network can be used to exactly compute any function. Rather, we can get an approximation that is as good as we want. By increasing the number of hidden neurons we can improve the approximation.
- tsbinz 7y agoAlso, the article doesn't mention a crucial part of the universal approximation theorem: It is about functions on compact subsets of R^n, so it doesn't say anything about functions that take the whole of R as input.
- ska 7y agoIt's also important to note the theorem doesn't help you find the right weights, just notes they exist.
- pauljurczak 7y agoThey can't approximate functions on unbound domain in general case.