About

This is a code for a paper "ArgDiver: Generating Sentential Arguments from Diverse Perspectives on Controversial Topics", accepted in NLP4IF 2019. If you meet any problem in our code, please feel free to use the Issues section in this repository.

Getting Started

Requirements

• tensorflow-gpu==1.8
• numpy==1.16.2
• stanfordcorenlp-3.9.1.1

Prerequistes

• Prepare the PERSPECTRUM dataset proposed in this paper.

• (Optional) Download the Wikitext-103 from here, proposed in this paper. You can use `download_wikitext.sh' script.

  • If you don't pre-train a model, just skip this part and use --use_pretrain=False flag when train a model.

• Download the Glove word embedding from here.

• After download the above resources, locate them in appropriate directory with following the below Package Structure.

Commands

1. Preprocessing

python preprocessing.py

(Optional) Pretrain the weights of encoder and decoder

python main.py --mode=lm_train --model=lm

2. Train model. If you don't pretrain the model in above, please set the --use_pretrain=False

python main.py --mode=train --model=[vanilla,embmin]

(Optional) To use MMI-bidi for decoding, train both the standard-seq2seq(vanilla) and the reverse-seq2seq model using below command.

python main.py --mode=train --model=mmi_bidi

3. Decode using trained model.

python main.py --mode=decode --model=[vanilla,mmi_bidi,embmin] --beam_size=10

Package Structure

├── sentential_argument_generation
│     └── beamsearch.py
│     └── data_loader.py
│     └── preprocessing.py
│     └── utils.py
│     └── main.py
│     └── models/
│          └──── _init_.py
│          └──── attention.py
│          └──── basemodel.py
│          └──── emb_min.py
│          └──── lm.py
│     └── data/
│          └──── emb/
│                 └──── glove.6B.300d.txt
│          └──── log/
│          └──── perspectrum/
│                      └──── (Locate the json files of PERSPECTRUM data to here!)
│          └──── wikitext/wikitext-103/
│                      └──── (Locate the Wikitext-103 .tokens files to here!)
│          └──── trainable/
│                      └──── split/

Results

Input
- We should fear the power of government over the internet.

Outputs:
- National sovereignty would result in a government’s freedom of expression.
- The government should not be celebrated.
- It is a necessary for national security.
- It’s conceivable to the wrong hands.
- The government is a best way to have a universal right to have a universal right to practice.

References

• The implementation of attention mechanism and neural network is based on this and this repositories.
• Chen, Sihao, et al. "Seeing Things from a Different Angle: Discovering Diverse Perspectives about Claims." Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). 2019. paper

Citation

@inproceedings{park-etal-2019-generating,
    title = "Generating Sentential Arguments from Diverse Perspectives on Controversial Topic",
    author = "Park, ChaeHun  and
      Yang, Wonsuk  and
      Park, Jong",
    booktitle = "Proceedings of the Second Workshop on Natural Language Processing for Internet Freedom: Censorship, Disinformation, and Propaganda",
    month = nov,
    year = "2019",
    address = "Hong Kong, China",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/D19-5007",
    doi = "10.18653/v1/D19-5007",
    pages = "56--65",
    abstract = "Considering diverse aspects of an argumentative issue is an essential step for mitigating a biased opinion and making reasonable decisions. A related generation model can produce flexible results that cover a wide range of topics, compared to the retrieval-based method that may show unstable performance for unseen data. In this paper, we study the problem of generating sentential arguments from multiple perspectives, and propose a neural method to address this problem. Our model, ArgDiver (Argument generation model from diverse perspectives), in a way a conversational system, successfully generates high-quality sentential arguments. At the same time, the automatically generated arguments by our model show a higher diversity than those generated by any other baseline models. We believe that our work provides evidence for the potential of a good generation model in providing diverse perspectives on a controversial topic.",
}