Uni-MIS: United Multiple Intent Spoken Language Understanding via Multi-View Intent-Slot Interaction

Official code repository for the paper Uni-MIS: United Multiple Intent Spoken Language Understanding via Multi-View Intent-Slot Interaction accepted at AAAI24.

Setup

In a conda env with pytorch / cuda available, run:

pip install -r requirements.txt

Train

Example script:

python train.py -g -hv -wt tf --decoder agif  --embed_type roberta 

Optimization Recommendations

We recommend performing hyperparameter optimization by using commands :

python wandb_atis.py 
python wandb_snips.py

These scripts are designed to search for the optimal parameters to enhance the model's performance.

##If you use our data or code, please cite our paper:

@inproceedings{yin2024uni,
    title={Uni-MIS: United Multiple Intent Spoken Language Understanding via Multi-View Intent-Slot Interaction},
    author={Yin, Shangjian and Huang, Peijie and Xu, Yuhong},
    booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
    volume={38},
    number={17},
    pages={19395--19403},
    year={2024}
    }