Architecture of the proposed MCP-NER. NER is modeled as a generation task, and the span position of the entity and the corresponding entity type are generated through the vector output by the decoder at each moment. The soft cue vector is added to the encoding and decoding process of BART to guide the attention flow in the transformer.
This site uses cookies. By continuing to use our website, you are agreeing to our privacy policy. No content on this site may be used to train artificial intelligence systems without permission in writing from the MIT Press.