HIGHLIGHTS
SUMMARY
Relationship extraction is a high-level task in the NLP field, and the dataset for this task requires annotation of both entities and relationships between entities, which further increases the workload of the medical relationship extraction annotation task. The introduction of subsequence annotation not only improves the efficiency of annotation, but also improves the performance of the relational extraction model. The authors propose a subsequence and distant supervision based active learning (SDSAL) method for Chinese medical texts relation extraction and experiment shows it could be beneficial for annotation efficiency by selecting information-rich subsequences . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.