Skip to main content

Submissions

Submisison are now open and due on Feburary 13th, 2023

Please make submissions directly in our open review: SLU Grand Challenge ICASSP Submission Link

Below we describe the submission format.

All tracks are required to submit a 2-page paper describing the method used to achieve the results following the ICASSP Paper Guidelines.

Additionally please share a single .zip file containing model summary and predictions on the test set as supplemental material, descriptions for each track are listed below.

Quality Track:

  • quality_model_summary.txt: Use Model Summary Format
  • quality_model_predictions.txt: All model predictions on overall test set, use Model Predictions Format

On-device Track:

  • ondevice_model_summary.txt: Use Model Summary Format
  • ondevice_model_predictions.txt: All model predictions on overall test set, use Model Predictions Format

Low Resource Track:

  • lr_weather_model_summary.txt: Model summary of the model used for the weather domain, Use Model Summary Format
  • lr_reminder_model_summary.txt: Model summary of the model used for the reminder domain, Use Model Summary Format
  • lr_weather_predictions.txt: All model predictions on weather test set, use Model Predictions Format
  • lr_reminder_predictions.txt: All model predictions on reminder test set, use Model Predictions Format

Model Summary Format: Reported EM and File Size as a text file, an indication whether the techniques used are open source, and a team name. Model Predictions Format: A TSV file containing, file_id, and model prediction as a string