Examining and Adapting Time for Multilingual Classification via Mixture of Temporal Experts

Table of Contents

• Environment Setup
• Usage
• Contact and Citation

Environment Setup

1. Platform:

• Ubuntu 22.04
• Anaconda, Python 3.10.13
• Linux Kernel: 6.8.0-40-generic

2. Run the following commands to create the environment:

• conda env create -f environment.yml
• conda activate tempo0

Usage

• Run sh train.sh to get the performance and checkpoints of the base model. You can change to your own choice of model and dataset in the sh file.
• Run sh mote.sh to adapt the base model.

Contact and Citation

For any inquiries, feel free to contact the author at: wliu9@memphis.edu

To cite this work:

@misc{liu2025examiningadaptingtimemultilingual,
      title={Examining and Adapting Time for Multilingual Classification via Mixture of Temporal Experts}, 
      author={Weisi Liu and Guangzeng Han and Xiaolei Huang},
      year={2025},
      eprint={2502.08825},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.08825}, 
}