UiPath UIPATH-SAIV1 Online Practice Questions and Exam Preparation

UiPath UIPATH-SAIV1 Online Questions & Answers

• Question 71:

  What is a possible outcome specific to the Low confidence training mode in the "Explore" phase while working with UiPath Communications Mining?

  A. The dataset will be enriched with additional data sources, so the entity extraction process may suffer significant changes.
  B. Due to the fact that most of the verbatims don't have any predictions, the training process may reveal the need to add new labels that are not yet covered in the taxonomy.
  C. Due to the fact that most labels have a high recall percentage, the training process may reveal the need for a "human in the loop" scenario.
  D. The pinned model may suffer from performance degradation, so the training process may require additional effort on a regular basis.
  B. Due to the fact that most of the verbatims don't have any predictions, the training process may reveal the need to add new labels that are not yet covered in the taxonomy.

• Question 72:

  When creating a training dataset, what is the recommended number of samples for a Column field?

  A. 10-20 document samples per column field.
  B. 20-50 document samples per column field.
  C. 50-200 document samples per column field.
  D. 100-200 document samples per column field.
  C. 50-200 document samples per column field.

• Question 73:

  Which of the following OCR (Optical Character Recognition) engines is not free of charge?

  A. Tesseract.
  B. Microsoft Azure OCR.
  C. OmniPage.
  D. Microsoft OCR.
  B. Microsoft Azure OCR.

• Question 74:

  What can be found in the Config file specific to the Document Understanding Process?

  A. Settings and configurations related to post-processing rules like field confidences.
  B. The labels for the Machine Learning model.
  C. The schema for the documents in Data Manager.
  D. Definitions for the input document types and formats supported by the Document Understanding Process.
  A. Settings and configurations related to post-processing rules like field confidences.

• Question 75:

  In UiPath Communications Mining, when creating a model that is intended to enable automation, what is the minimum F1 score recommended for entities?

  A. 60%
  B. 70%
  C. 80%
  D. 90% and above
  D. 90% and above

• Question 76:

  Which is the most suitable extractor for extracting data from invoices from different customers?

  A. The Regex Based Extractor.
  B. The Machine Learning Extractor.
  C. The Form Extractor.
  D. The Intelligent Form Extractor.
  B. The Machine Learning Extractor.

• Question 77:

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

  A. Using a Time Trigger that runs every 5 minutes.
  B. Using a Queue Trigger for parallel jobs, tasked to start a new job for every new Queue item.
  C. Having the Dispatcher process use a Start Job activity.
  D. Starting a batch job that runs the Document Understanding Process for all the needed documents.
  B. Using a Queue Trigger for parallel jobs, tasked to start a new job for every new Queue item.

• Question 78:

  Which of the following statements is correct regarding ML (Machine Learning) model best practices?

  A. Building an ML model doesn't require feature engineering. The best practice is simply inputting the raw data into the model for obtaining accurate predictions.
  B. The best practice when building an ML model is to use as much data as possible, regardless of its quality or diversity. More data always leads to better model performance, even if the data is unbalanced or contains errors because the model has much more information it can learn from.
  C. When building an ML model, it is best to train it on the entire dataset without splitting it into training, validation, and test sets. This approach allows the model to learn from all available data, resulting in optimal performance.
  D. Collecting high-quality and diverse data is a best practice when building an ML model. This includes ensuring data cleanliness, addressing class imbalance if present, and validating data integrity. Splitting the entire dataset into training, validation, and test is also recommended.
  D. Collecting high-quality and diverse data is a best practice when building an ML model. This includes ensuring data cleanliness, addressing class imbalance if present, and validating data integrity. Splitting the entire dataset into training, validation, and test is also recommended.

• Question 79:

  What action can be performed when using the User Interface to update a UiPath Communications Mining entity?

  A. Edit entity hierarchy.
  B. Clone entity.
  C. Change base entity.
  D. Merge with another entity.
  C. Change base entity.

• Question 80:

  Which are all the options for managing ML Skills?

  A. ML skills can be created, stopped, redeployed, updated to a new package version, rolled back to a previous package version, modified to use or not use GPU, made public or private, or deleted.
  B. ML skills can be created, stopped, redeployed, updated to a new package version, rolled back to a previous package version, made public or private, or deleted.
  C. ML skills can be created, stopped, redeployed, updated to a new package version, rolled back to a previous package version, modified to use or not use GPU, modified to use or not use AI units, made public or private, or deleted.
  D. ML skills can be created, updated to a new package version, rolled back to a previous package version, modified to use or not use GPU, made public or private, or deleted.
  A. ML skills can be created, stopped, redeployed, updated to a new package version, rolled back to a previous package version, modified to use or not use GPU, made public or private, or deleted.