gradient with hypernetworks following STN [8], which adds a linear transformation between hyperparameters and

Neural Information Processing Systems 

A more detailed computational analysis will be added in the final version. We also thank the reviewer for pointing out the valuable reference. HPM on CIFAR-10 are 78.5/78.1, We will include them in the final version. HPM, GB-HPO + RS may not fully explore the hyperparameter space due to the lack of mutation-driven search.

Similar Docs  Excel Report  more

TitleSimilaritySource
None found