Fine-grained Text Sentiment Transfer

This repository contains the original implementation of the models presented in Towards Fine-grained Text Sentiment Transfer (ACL 2019).

Reproducibility

In order to help you quickly reproduce the existing works of fine-grained text sentiment transfer, we release the outputs of all models and the corresponding references.

• Outputs: Generated results (outputs) of 3 baselines and our model are in the outputs/ directory.
• References: Human references are in the data/yelp/reference.txt file.

Dependencies

python==2.7
numpy==1.14.2
tensorflow==1.13.1
OpenNMT-tf==1.15.0 

Quick Start

Step 1: Pre-train the sentiment scorer

A pretrained sentiment scorer is used to compute the sentiment transformation reward. Here the scorer is implemented as LSTM-based linear regression model. You can train the model using the following command:

cd regressor/
python main.py --mode train

Note: If you get the error no module named opennmt, please install OpenNMT-tf: pip install OpenNMT-tf==1.15.0.

Step 2: Pre-train the Seq2SentiSeq model using pseudo-parallel data

You can train the Seq2SentiSeq model using the following command:

cd seq2sentiseq/
python main.py --mode train

Step 3: Cycle reinforcement learning

After finishing the previous two steps, you can start the cycle reinforcement learning using the following command:

python cycle_training.py --n_epoch 30

The final transffered results are in the ../tmp/output/yelp_final_*/ dir.

Cite

Please cite the following paper if you found the resources in this repository useful.

@inproceedings{luo2019towards,
  title={Towards Fine-grained Text Sentiment Transfer},
  author={Luo, Fuli and Li, Peng and Yang, Pengcheng and Zhou, Jie and Tan, Yutong and Chang, Baobao and Sui, Zhifang and Sun, Xu},
  booktitle={Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, {ACL} 2019},
  pages={2013--2022},
  year={2019}
}