Title | ||
---|---|---|
Encoding multi-granularity structural information for joint Chinese word segmentation and POS tagging |
Abstract | ||
---|---|---|
•We are the first that improve the joint Chinese word segmentation and POS tagging, by using multi-granularity structural information.•We construct information graph based on the character, word and subword, and encode them via lattice-LSTM and GCN model.•We obtain the new best performances on the five benchmarks for the joint task, and also conduct in-depth analysis.•Our methods also can help to relieve the out-of-vocabulary and the long-range dependency issues for the joint tasks. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1016/j.patrec.2020.07.017 | Pattern Recognition Letters |
Keywords | DocType | Volume |
Chinese word segmentation,POS tagging,Joint model,Lattice model,Graph model | Journal | 138 |
ISSN | Citations | PageRank |
0167-8655 | 1 | 0.35 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ling Zhao | 1 | 13 | 9.23 |
Ailian Zhang | 2 | 1 | 0.35 |
Ying Liu | 3 | 1 | 0.35 |
Hao Fei | 4 | 16 | 15.51 |