AskNowQA / DeepShallowParsingQA

MDP-Parser, A shallow parser for KGQA

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

MDP-Parser

The main objective of this project is to find the mention of entities and relation given a natural language question and underlying knowledge graph. It further use a simple index-based linking to link the mentions into the item from the knowledge graph.

Train

Parameters --epochs --dropout --lr --batchsize --k --mode (train/test) --dataset (lcquad/qald_7_ml) --b (number of prev/next words) --checkpoint filename --policy (lstm/bilstm/nn)

Sample run command:

python scripts/execute.py --epochs 50  --dropout 0.1 --lr 0.0001 --batchsize 1 --k 10 --mode train --dataset qald_7_ml --b 1 --checkpoint tmp --policy nn

Cite

If you would use the codebase, please cite our paper as well:

@article{zafar2019mdp,
  title={Distantly Supervised Question Parsing},
  author={Zafar, Hamid and Tavakol, Maryam and Lehmann, Jens},
  booktitle={European Conference on Artificial Intelligence},
  year={2020}
}

Live Demo

Coming soon! Stay tuned!

License

This program is subject to the terms of the Apache License Version 2.0.

About

MDP-Parser, A shallow parser for KGQA

License:Apache License 2.0


Languages

Language:Python 99.6%Language:Dockerfile 0.4%