发帖
雷达卡
7516
7517
Active Learning (AL): The traditional AL methods are as softmax output of the classifier, hθ (x) is a composition
iterative schemes which obtain labels (from oracle or of fθ gθ (x) where, gθ (x) is the function that maps input
experts) for a set of informative data samples. The newly to embedding and fθ does final classification. The domain
labeled samples are then added to the pool of existing discriminator is a network Dφ (gθ (x)) which classifies the
lab ...
京公网安备 11010802022788号
