UiSAIv1

UiSAIv1 practice questions, page 18 of 20, with answer review and explanations.

Question 86

When adding a new label to a mature taxonomy in UiPath Communications Mining, why is it important to apply it to all of the relevant examples in the previously reviewed set?

  1. Training data for each label should always be evenly spread across the time period covered in the dataset.
  2. If the new label is not applied to relevant previously reviewed verbatims, the model will start predicting other labels less accurately.
  3. It will reduce bias for the new label.
  4. The previously reviewed examples where the new label should also have been applied will cause confusion for the model if not corrected.

Question 87

What is the correct way to start a Document Understanding job in Unattended scenarios?

  1. Having the Dispatcher process use a Start Job and Get Reference activity.
  2. Starting a batch job that runs the Document Understanding Process for all the needed documents.
  3. Starting the job from Orchestrator from the Processes tab.
  4. Using a Time Trigger that runs every 5 minutes.

Question 88

What is the correct approach when discovering that a label's name does not reflect the desired intent within a UiPath Communications Mining taxonomy?

  1. Export dashboards where the original label appeared, then manually change the name.
  2. Create another label with the correct name and apply it to all verbatims that have the original label.
  3. Create a label intending to show the error in naming and insert it into all dashboards where the original label appeared.
  4. Rename the label in the taxonomy as it doesn't negatively affect the model training and performance.

Question 89

What is the primary goal of exception training for a production model in UiPath Communications Mining?

  1. To teach the model what to ignore in the future.
  2. To improve the model's ability to confidently predict future similar verbatims.
  3. To identify user interface issues.
  4. To remove examples from the dataset that the model could not confidently make predictions for.

Question 90

In an analytics-focused model in UiPath Communications Mining, which factor determines the extent to which a model trainer should optimize their model's performance?

  1. The trainer's discretion is based on use case objectives and capacity to continue training.
  2. There are at least 200 labels with at least a 70% MAP.
  3. There are at least 150 labels with at least an 80% MAP.
  4. The proportion of data that has been annotated is at least 10% of total dataset volumes.