Fusing Part-of-Speech Information in Low-Resource Neural Paraphrase Generation

Figure 4

The “dc” encoder.