chocowu / diare-d2g

Codes of of the IJCAI 2022 Paper Global Inference with Explicit Syntactic and Discourse Structures for Dialogue-Level Relation Extraction

Python implementation of IJCAI 2022 Paper Global Inference with Explicit Syntactic and Discourse Structures for Dialogue-Level Relation Extraction

---

Requirement

Install the packages as in requirement.txt.

pip install -r requirement.txt

Data Preprocessing

• Downloading the the dataset and pretrained embeddings, which are officially provided by Dialogue-Based Relation Extraction.
• Constructing Dialogue-level Mixed Dependency Graph for instances via data/D2G_construction.
• Preparing data inputs via data/gen_data.py.

System Running:

• Training:

python train.py \
        --save_name dialogre  \
        --use_spemb True  \
        --use_wratt True  \
        --use_arc True

• Testing:

python test.py \
         --save_name dialogre \
         --use_spemb True \
         --use_wratt True  \
         --use_arc True

Credit

Codes are based on the repos of 1) the ACL-20 Paper "Reasoning with Latent Structure Refinement for Document-Level Relation Extraction", and TASLP Paper "Relation Extraction in Dialogues: A Deep Learning Model Based on the Generality and Specialty of Dialogue Text".

Citation

@inproceedings{FeiDiaREIJCAI22,
  title     = {Global Inference with Explicit Syntactic and Discourse Structures for Dialogue-Level Relation Extraction},
  author    = {Fei, Hao and Li, Jingye and Wu, Shengqiong and Li, Chenliang and Ji, Donghong and Li, Fei},
  booktitle = {Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, {IJCAI}},
  pages     = {4082--4088},
  year      = {2022},
}