xiaoanshi / physical-activity-counseling

Dataset and Codebase for Evaluation of In-Person Counseling Strategies To Develop Physical Activity Chatbot for Women, published as a long paper in SIGDIAL 2021.

Evaluation of In-Person Counseling Strategies To Develop Physical Activity Chatbot for Women

Dataset and Codebase for Evaluation of In-Person Counseling Strategies To Develop Physical Activity Chatbot for Women, published as a long paper in SIGDIAL 2021.

Citation

If you would like to refer to our work, please cite the following BibTex entry:

  @inproceedings{liang2021evaluation,
  author={Kai-Hui Liang, Patrick Lange, Yoo Jung Oh, Jingwen Zhang, Yoshimi Fukuoka, Zhou Yu},
  title={Evaluation of In-Person Counseling Strategies To Develop Physical Activity Chatbot for Women},
  journal={Proceedings of the 22nd Annual Meeting of the Special Interest Group on Discourse and Dialogue},
  pages={32--44},
  year={2021}
  }

Dataset

The simulated dialog is under directory data/.
Since releasing the original interview data is not approved by our IRB and HIPPA, we created 44 dialogs (772 sentences) based on the original interview daialog for our community to use.

Annotation Scheme

The annotation scheme guidelines can be found under Annotation Scheme.pdf.

Strategy Classifier

The code of the strategy classifier is under directory classifier/.

To use the fine-tuned model for strategy prediction, please download training_args.binfrom here and optimizer.pt from here and put them under the classifier/models/strategyfolder.

More instructions are provided in the classifier/README.md file.