Skip to main content

Table 2 Parameters design

From: Incorporating representation learning and multihead attention to improve biomedical cross-sentence n-ary relation extraction

Parameter name Value
Word embedding dimension 200
Subrelation embedding dimension 50
Position embedding dimension 50
Recurrent dropout for Bi-LSTM 0.5
GCN dropout probability 0.5
Batch size 6
Adam-learning rate 0.001
Hidden state dimension of Bi-LSTM 200
Hidden state dimension of multihead 400
Multihead attention head 4