Machine Learning is Fun! Part 3: Deep Learning and Convolutional Neural Networks
Adam Geitgey

Would anything improve if you took the false positives and false negatives and re-trained using them, sort of “correcting” the mistakes that were made? I’m guessing that would only really help if the mistakes followed a particular pattern, e.g. the network always fails at recognizing birds on a blue background

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.