Climate-NLI

This repo contains the files for training and evaluating Climate-NLI, a model designed for fact-checking and zero-shot classification on climate-related text.

How to run

To run the experiment, simply use python command followed by run_*.py, based on the experiment which you want to run along with the arguments.

Run the Climate-NLI model training

python run_nli_train.py --add_label_variation

Run the Climate-NLI model evaluation

The available model name are :

• bart-large-mnli
• climate-nli-binary
• climate-nli-unseen-binary

or you can defined the name by yourself here and set the condition on ./src/eval/flow.py.

python run_nli_train.py --autolabel_type <model_name>

Citation

Our work was presented at PACLIC38

@inproceedings{yudanto-etal-2024-climate,
    title = "Climate-{NLI}: A Model for Natural Language Inference and Zero-Shot Classification on Climate-Related Text",
    author = "Yudanto, Faturahman  and
      Sari, Yunita  and
      Zaki, Maeve Zahwa Adriana Crown",
    editor = "Oco, Nathaniel  and
      Dita, Shirley N.  and
      Borlongan, Ariane Macalinga  and
      Kim, Jong-Bok",
    booktitle = "Proceedings of the 38th Pacific Asia Conference on Language, Information and Computation",
    month = dec,
    year = "2024",
    address = "Tokyo, Japan",
    publisher = "Tokyo University of Foreign Studies",
    url = "https://aclanthology.org/2024.paclic-1.57/",
    pages = "600--608"
}