Abstract: We propose a new method for domain adaptation that uses a gradient reversal layer (GRL) with instance weighting. Domain adaptation methods that use GRL have the latent problem of learning ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results