H2DT

This repository contains data and code for the AAAI 2024 paper: Harnessing Holistic Discourse Features and Triadic Interaction for Sentiment Quadruple Extraction in Dialogues

---

To clone the repository, please run the following command:

git clone https://github.com/whulacc/h2dt

Quick Links

• Overview
• Requirements
• Code Usage
• Citation

Overview

In this paper, we propose a new model, named H2DT, to extract Target-Aspect-Opinion-Sentiment quadruples from the given dialogue. More details about the task can be found in our paper.

Requirements

This project uses PyTorch for its implementation. Please ensure your system includes the following package versions:

• Python: 3.7+
• PyTorch: 1.13.1+

Additional required packages can be installed via pip:

pip install -r requirements.txt

You should install dgl with the following command:

pip install dgl -f https://data.dgl.ai/wheels/cu117/repo.html

Code Usage

For Training and Evaluating on the Chinese Dataset

Run the following script to train and evaluate:

bash scripts/train_zh.sh

For Training and Evaluating on the English Dataset

Run the following script to train and evaluate:

bash scripts/train_en.sh

GPU Memory Requirements

Below is a table detailing the GPU memory requirements based on dataset and batch size:

Dataset	Batch Size	GPU Memory Required
Chinese	2	20GB
English	1	20GB

Customizing Hyperparameters

Hyperparameter settings are flexible and can be adjusted within either main.py or src/config.yaml. Note that configurations in main.py will override any settings in src/config.yaml.

Citation

If you would like to cite our work, please use the following format:

@inproceedings{lih2dt-aaai-24,
  author       = {Bobo Li and Hao Fei and Lizi Liao and Yu Zhao and Fangfang Su and Fei Li and Donghong Ji},
  title        = {Harnessing Holistic Discourse Features and Triadic Interaction for
                  Sentiment Quadruple Extraction in Dialogues},
  booktitle    = {AAAI},
  pages        = {18462--18470},
  year         = {2024},
}