xwhan / Knowledge Aware Reader

Code for the ACL 2019 paper:

Improving Question Answering over Incomplete KBs with Knowledge-Aware Reader

Paper link: https://arxiv.org/abs/1905.07098

Model Overview:

Requirements

• PyTorch 1.0.1
• tensorboardX
• tqdm
• gluonnlp

Prepare data

mkdir datasets && cd datasets && wget https://sites.cs.ucsb.edu/~xwhan/datasets/webqsp.tar.gz && tar -xzvf webqsp.tar.gz && cd ..

Full KB setting

CUDA_VISIBLE_DEVICES=0 python train.py --model_id KAReader_full_kb --max_num_neighbors 50 --label_smooth 0.1 --data_folder datasets/webqsp/full/ 

Incomplete KB setting

Note: The [email protected] should match or be slightly better than the number reported in the paper. More tuning on threshold should give you better F1 score.

30% KB

CUDA_VISIBLE_DEVICES=0 python train.py --model_id KAReader_kb_03 --max_num_neighbors 50 --use_doc --data_folder datasets/webqsp/kb_03/ --eps 0.05

10% KB

CUDA_VISIBLE_DEVICES=0 python train.py --model_id KAReader_kb_01 --max_num_neighbors 50 --use_doc --data_folder datasets/webqsp/kb_01/ --eps 0.05

50% KB

CUDA_VISIBLE_DEVICES=0 python train.py --model_id KAReader_kb_05 --num_layer 1 --max_num_neighbors 100 --use_doc --data_folder datasets/webqsp/kb_05/ --eps 0.05 --seed 3 --hidden_drop 0.05

Citation

@inproceedings{xiong-etal-2019-improving,
    title = "Improving Question Answering over Incomplete {KB}s with Knowledge-Aware Reader",
    author = "Xiong, Wenhan  and
      Yu, Mo  and
      Chang, Shiyu  and
      Guo, Xiaoxiao  and
      Wang, William Yang",
    booktitle = "Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics",
    month = jul,
    year = "2019",
    address = "Florence, Italy",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/P19-1417",
    doi = "10.18653/v1/P19-1417",
    pages = "4258--4264",
}