And how much training data does it need to achieve high accuracy?
QR scanning is actually a non-trivial algorithm. If NN can learn this algorithm, is it possible to reverse engineer any complicated software system, given enough training data (input/output pairs)?
QR scanning is actually a non-trivial algorithm. If NN can learn this algorithm, is it possible to reverse engineer any complicated software system, given enough training data (input/output pairs)?
Thoughts?