Mutual-Attention Net: A Deep Attentional Neural Network for Keyphrase Generation

Table 3

Comparison results of predicting the absent keyphrases on recall at top 10 and top 50.

	RNN		CopyRNN		Our model
	R@10	R@50	R@10	R@50	R@10	R@50
Inspec	0.031	0.061	0.047	0.100	0.051	0.108
Krapivin	0.095	0.156	0.113	0.202	0.118	0.206
NUS	0.050	0.089	0.058	0.116	0.064	0.119
SemEval	0.041	0.060	0.043	0.067	0.046	0.084
KP20k	0.083	0.144	0.125	0.211	0.148	0.227

The best results are highlighted in bold.