PMI Cognitive Project Management in AI (PMI-CPMAI) - CPMAI Exam Practice Test

A. Data leakage B. Bias elimination C. Underfitting D. Model instability

A. Using a hybrid encryption scheme for storage B. Utilizing a secure multiparty computation framework C. Implementing a zero-trust architecture for network security D. Applying data anonymization to the dataset

A. Developing a comprehensive data anonymization strategy B. Implementing a minimalistic data collection policy C. Ensuring regular impact assessments are conducted to meet privacy laws D. Establishing robust data encryption protocol for privacy

A. Implement just-in-time inventory B. Establish contingency plans C. Engage with multiple suppliers D. Increase resource allocation

A. Attended bot B. Unattended bot C. Augmented Intelligence D. RPA is not the right solution to this problem

A. Infrastructure latency in the serving environment B. Insufficient model complexity C. Concept drift in the relationship between features and outcomes D. Data leakage in the original training set

A. Create a data visualization. B. Understand the data characteristics. C. Evaluate the data freshness and relevance. D. Delete the suspicious data manually.

A. Phase II B. Phase IV C. Phase V D. Phase VI E. Phase I F. Phase III

A. Move forward with the project as planned. The problem definition will become clear later on in the project. B. Do not move forward until you can determine a clear problem definition. C. Do not move forward and cancel the project all together. D. Cautiously move forward as planned. You do not need to answer yes to all the questions in the AI Go/No Go decision chart to start your project.

A. Conducting a series of penetration tests B. Testing against validation datasets C. Implementing an impact evaluation D. Running multiple end-user acceptance tests

A. False Negative results B. True Positive results C. False Positive results D. True Negative results

A. Selecting a simple algorithm such as a Gaussian mixture model B. Selecting a simple algorithm such as Naive Bayes C. Selecting an ensemble method if uncertain as to which algorithm provides the best performance D. Selecting a known neural network algorithm that works well for supervised learning approaches

A. User satisfaction, containment rate, escalation accuracy, and privacy/compliance incidents B. Response time only C. Lines of code written D. Number of features delivered

A. Implement a batch processing system to enhance performance. B. Adopt a rule-based extraction, transformation, and loading (ETL) framework. C. Utilize an advanced data distribution service (DDS). D. Employ machine learning (ML) algorithms.