Multipath Cross Graph Convolution for Knowledge Representation Learning

Table 2

The optimal parameter configuration of the corresponding datasets.

Parameters	Datasets
	YAGO43kET	WN18RR	FB15kET
	0.1	0.1	0.1
	0.6	0.5	0.4
	200	150	250
	100	100	125
Batch size	4096	4096	4096
Epochs	300	300	300