AMR-Backparsing

An implementation for paper "Online Back-Parsing for AMR-To-Text Generation" (in EMNLP 2020)

Requirements

• python 3.6
• pytorch 1.0

Data Preprocessing

We follow this work to preprocess AMR. Since AMR corpus require LDC license, we upload some examples for format reference. If you have the license, feel free to contact us for getting the preprocessed data.

Runing

bash ./src/train-LDC2015.sh
bash ./src/train-LDC2017.sh

Evaluation

bash ./src/translate-LDC15.sh
bash ./src/translate-LDC17.sh

Results

Setting	BLEU-tok	BLEU-nltk	Meteor	chrF++
LDC15	31.58	32.27	36.38	65.33
LDC17	34.36	34.98	38.09	67.90

References

@inproceedings{bai-etal-2020-online,
    title = "Online Back-Parsing for {AMR}-to-Text Generation",
    author = "Bai, Xuefeng  and
      Song, Linfeng  and
      Zhang, Yue",
    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.emnlp-main.92",
    doi = "10.18653/v1/2020.emnlp-main.92",
    pages = "1206--1219",
}