For a semantic segmentation problem attempted using neural networks, does it make sense to try achieving overfitting with single training example and then (depending on generalization error) add more training examples and modify the neural network incrementally?

  • $\begingroup$ no, not really. $\endgroup$ – shimao Apr 10 '19 at 7:07
  • $\begingroup$ What would you like to achieve by this? $\endgroup$ – Jan Kukacka Apr 10 '19 at 8:11
  • $\begingroup$ Due to lack of training data(only 5 labeled examples right now), I was thinking of starting with only one training example and expand n/w based on generalization error. $\endgroup$ – Pranav Apr 10 '19 at 9:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.