ATOSS: Aspect Term-Oriented Sentence Splitter

Repo for paper (EMNLP 2024 Findings)

📄 Make Compound Sentences Simple to Analyze: Learning to Split Sentences for Aspect-based Sentiment Analysis

Quick Start

Follow the steps below ⬇️

1. Set Up Environment

conda create -n atoss python=3.8
conda activate atoss
pip install torch==1.13.0+cu117 torchvision==0.14.0+cu117 torchaudio==0.13.0 --extra-index-url https://download.pytorch.org/whl/cu117
pip install -r requirements.txt

2. Train & Inference

run with command:

cd src

Run SFT (Supervised Fine-Tuning):

sh run_sft.sh

Run DPO (Direct Preference Optimization):

sh run_dpo.sh

3. Evaluate Results

Once the inference steps for SFT and DPO are completed, you can evaluate the results using baseline models:

Fine-tuned models
- Paraphrase, DLO & ILO and MVP
Prompting-based LLMs
- GPT-3.5, GPT-4 and GPT-4o

Evaluation files:

sft_test.txt (for SFT results)
dpo_test.txt (for DPO results)

5. Citation

If you find ATOSS helpful for your research, please cite:

@article{seo2024make,
  title={Make Compound Sentences Simple to Analyze: Learning to Split Sentences for Aspect-based Sentiment Analysis},
  author={Seo, Yongsik and Song, Sungwon and Heo, Ryang and Kim, Jieyong and Lee, Dongha},
  journal={arXiv preprint arXiv:2410.02297},
  year={2024}
}

6. Contact

For questions, suggestions, or issues, feel free to reach out:

Email: ryang1119@yonsei.ac.kr

ryang1119/ATOSS