Neural Symbolic Number Reasoning for Natural Language Inference

AWPNLI task implementation for paper A Neural-Symbolic Approach to Natural Language Understanding

Usage

0. install the dependencies in requirement.txt
1. install the spacy small en language model.
2. split the labeled AWP dataset, see labeled_data/AWPNLI.jsonl

python3 src/split_dataset.py  \
        --input_data_file labeled_data/AWPNLI.jsonl  \
        --output_data_folder dataset/AWPNLI  \
        --mode=cv-10

3. run the specific model on the specific split by setting the config file in the config folder. e.g.

python3 src/launch_experiment.py -c $config_file --cuda $cuda

4. gather results by in the default output folder experiment

python3 src/log_reader.py --folder experiment

Step 2-4 can also be found in run.sh file

Support model

• {bart/roberta}_cls_3way: neural NLI classifier
• bart_forms_3way: symbolic execution
• shared_encoder_3way: neural symbolic with shared encoder

Bibliography

If our work inspires you, please cite

@inproceedings{liu2022a,
  title={A Neural-Symbolic Approach to Natural Language Understanding},
  author={Liu, Zhixuan and Wang, Zihao and Lin, Yuan and Li, Hang},
  booktitle={Findings of the Association for Computational Linguistics: EMNLP 2022},
  year={2022}
}