SECP

Code for Paper: Paraphrase Makes Perfect: Leveraging Expression Paraphrase to Improve Implicit Sentiment Learning

Overview

In this repository, we provide code for Sentiment Expression Conversion based Paraphrase (SECP), which focuses on improving implicit sentiment learning in Aspect-based Sentiment Analysis.

Requirements

• cuda 11.4
• Python 3.10.12
  • PyTorch==2.0.1
  • Transformers==4.33.0
  • scikit-learn====1.3.0
  • openprompt==1.0.1
  • PyYAML==6.0.1
  • numpy==1.25.2
  • sentencepiece==0.1.96

Datasets

To use the implicit_sentiment labeling provided by SCAPT-ABSA, we reformatted the data from SemEval2014 Laptop/Restaurant following ASGCN and appended a label to each sample, which indicates whether it is an implicit sentiment expression ("Y" and "N" indicate the implicit and explicit sentiment expression respectively). The data and paraphrased sentences for each datasets are provided in data and data/paraphrased_data.

Usage

# Create a virtual environment and install necessary packages.
bash setup.sh
# Activate the virtual environment.
source .venv/bin/activate

# Train the models according to the configs (./config)
bash run.sh

Checkpoints are saved in saved_checkpoints.

Citation

If you find this work useful, please cite the following:

@inproceedings{li-etal-2025-paraphrase,
    title = "Paraphrase Makes Perfect: Leveraging Expression Paraphrase to Improve Implicit Sentiment Learning",
    author = "Li, Xia  and
      Wang, Junlang  and
      Zheng, Yongqiang  and
      Chen, Yuan  and
      Zheng, Yangjia",
    editor = "Rambow, Owen  and
      Wanner, Leo  and
      Apidianaki, Marianna  and
      Al-Khalifa, Hend  and
      Eugenio, Barbara Di  and
      Schockaert, Steven",
    booktitle = "Proceedings of the 31st International Conference on Computational Linguistics",
    month = jan,
    year = "2025",
    address = "Abu Dhabi, UAE",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.coling-main.245/",
    pages = "3631--3647",
}