Hi, Thanks for your excellent work!
But I don't understand Eq.14. and Eq.15.:
why are the labels all-zero when the input is $$G_{\beta}\left(F_{\alpha_{s}}\right)$$ ?
Does all-zeros mean low confidence and all ones mean high confidence?
However, the $$G_{\beta}\left(F_{\alpha_{s}}\right)$$ is the network prediction. With the learning of network, is it similar to Y?
Hi, Thanks for your excellent work! But I don't understand Eq.14. and Eq.15.:
why are the labels all-zero when the input is
$$G_{\beta}\left(F_{\alpha_{s}}\right)$$? Does all-zeros mean low confidence and all ones mean high confidence? However, the$$G_{\beta}\left(F_{\alpha_{s}}\right)$$is the network prediction. With the learning of network, is it similar to Y?