This repo is the code and synthetic data of the ACM MM 2024 paper "SCREEN: A Benchmark for Situated Conversational Recommendation".
The organized dataset will be uploaded to Google Drive soon and you can download it from the Google Drive link.
The required packages are listed in requirements.txt. Suppose you use Anaconda to manage the Python dependencies, you can install them by running:
conda create -n screen python=3.11
conda activate screen
pip install -r requirements.txt
We will also upload the organized scene snapshot to Google Driver. Please put it in the scene_info_pool in the root directory.
Please set your openai key and other related parameters, and then run dialogue_simulation.py to start constructing data.
# set your OpenAI API key
export OPENAI_API_KEY=""
python dialogue_simulation.py
If you hope NOT to show the instructions and the synthesized conversations in the console, please set --show_description and --show_message to false.
Our code is partially based on the implementation of ChatArena. We thank the authors for their excellent work.
If you use our data or code in your work, please kindly cite our work as:
@inproceedings{lin-etal-2024-screen,
title = "SCREEN: A Benchmark for Situated Conversational Recommendation",
author = "Lin, Dongding and
Wang, Jian and
Leong, Chak Tou and
Li, Wenjie",
year = {2024},
isbn = {9798400706868},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3664647.3681651},
doi = {10.1145/3664647.3681651},
pages = {9591–9600},
numpages = {10},
keywords = {benchmark, role-playing, situated conversational recommendation},
location = {Melbourne VIC, Australia},
series = {MM '24}
}