3 ms·
Not really...It has limitations. Does not specify how large the neural network must be to achieve the desired level of approximation, does not address the compu
by belter 3y ago
Not really...It has limitations. Does not specify how large the neural network must be to achieve the desired level of approximation, does not address the computational resources required, ignores non continuous functions and so on...
"The Truth About the [Not So] Universal Approximation Theorem" - https://www.lifeiscomputation.com/the-truth-about-the-not-so-universal-approximation-theorem/ https://www.lifeiscomputation.com/the-truth-about-the-not-so...