Attention-Based Graph Convolutional Network for Zero-Shot Learning with Pre-Training

Algorithm 1

AGCNZ algorithm.

	Input: Adjacency matrix A, Number of nodes N, Input node features X, Pretrained ResNet50 model classifier parameters
	Output: Classifier parameter , Predicted categories of Unseen classes .
(1)	Initializes: the graph convolutional network parameters.
(2)	while not converged do
(3)	Update by equation (4);
(4)	for Attention-layer do
(5)	Update by equation (7);
(6)	Update by equation (5);
(7)	end for
(8)	Loss = LossFunction (, ), LossFunction update by equation (2) or (3);
(9)	Loss.backward;
(10)	end while
(11)	return
(12)	is obtained by using as classifier parameter of classification .