Skip to content

Experiment File

Experiment File for the Interaction Annotation UI

  • “exp_id” - id of the experiment.
  • “description” - description of the experiment to run.
  • “tasks”- listofassessments.
  • “query_title” - the title of the query to be displayed.
  • “setting” - an extra requirement that the user should take into account when doing the assessment.
  • “ranking_type” - the order in which to display the ranking of the documents. It should match the“ranking_type” defined in the configuration file. “ranking_type” could have the following values:
    • “original” - sorted by the “score”.
    • “ranker_” - sorted by the predicted score of a pre-trained model.
    • “pre/in/postprocessing_” - displays the ranking generated by the fairness intervention.

Experiment File for the Score Annotation UI

  • “exp_id” - id of the experiment.
  • “description” - description of the experiment to run.
  • “tasks”- listofassessments.
  • “query_title” - the title of the query to be displayed.
  • “setting” - an extra requirement that the user should take into account when doing the assessment.
  • “index” - the index of the item to be annotated.

Experiment File for the Ranking Comparison UI

  • “exp_id” - id of the experiment.
  • “description” - description of the experiment to run.
  • “tasks” - list of assessments.
  • “query_title” - the title of the query to be displayed.
  • “ranking_type_1” - the order in which to display the ranking of the documents that is displayed on the left side. It should match the “ranking_type” defined in the configuration file.
  • “ranking_type_2” - the order in which to display the ranking of the documents that is displayed on the right side. It should match the “ranking_type” defined in the configuration file.
    • “ranking_type” could have the following values:
    • “original” - sorted by the “score”.
    • “ranker_” - sorted by the predicted score of a pre-trained model.
    • “pre/in/postprocessing_” - displays the ranking generated by the fairness intervention.