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

Figure 3

The “cat” encoder.