Hacker News new | ask | show | jobs
by scollet 1497 days ago
Don't most production NN or DLN optimize to a maximum?

Seems like the behavior becomes predictable and then you have to retrain if you see unoptimal results.