Free USAII CAIC Practice Exam with Questions & Answers

The correct answer is D. All of the above because robotics, game playing, and navigation are all common examples of reinforcement learning. Reinforcement learning is a machine learning approach in which an agent learns by interacting with an environment and receiving rewards or penalties based on its actions. Over time, the agent learns a policy that helps it maximize long-term reward.

Robotics is a strong example because robots can learn movement, object handling, path planning, and control actions through trial and feedback. Game playing is another classic reinforcement learning example because an AI agent can learn winning strategies by trying actions, observing outcomes, and improving decisions over repeated episodes. Navigation is also a valid example because an agent can learn the best route or movement strategy by receiving feedback about distance, obstacles, time, or success in reaching a goal.

Since all three listed options are valid applications of reinforcement learning, the correct answer is D. All of the above .

The correct answer is D. Semi Reinforcement learning because it is not commonly recognized as a standard learning category for machine learning models. The major machine learning categories include supervised learning, unsupervised learning, reinforcement learning, and semi-supervised learning. Supervised learning uses labeled datasets where the model learns from known input-output examples. Unsupervised learning uses unlabeled data to discover patterns, clusters, or hidden structures. Reinforcement learning trains an agent through interaction with an environment using rewards and penalties. Semi-supervised learning combines a small amount of labeled data with a larger amount of unlabeled data to improve learning when fully labeled datasets are limited.

“Semi Reinforcement learning” is not normally listed as a core ML learning category in standard AI and machine learning learning paths. Therefore, among the given options, the one that is NOT a learning category for the ML model is D. Semi Reinforcement learning .

The correct answer is C. K-nearest neighbors KNNs because KNN is commonly used as a supervised learning algorithm, not an unsupervised learning algorithm. In supervised learning, the model uses labeled data to classify or predict outcomes for new data points. KNN works by comparing a new data point with nearby labeled examples and assigning a class or value based on those neighbors.

K-means clustering is a common unsupervised learning algorithm because it groups unlabeled data into clusters based on similarity. Principal Component Analysis PCA is also commonly associated with unsupervised learning because it reduces data dimensions by finding important patterns or directions of variance without requiring labeled outputs.

Since options A and B are valid unsupervised learning techniques, they are not the answer. The option that is not a correct common unsupervised learning model or algorithm is C. K-nearest neighbors KNNs .

The correct answer is A. supervised learning because supervised learning uses labeled data to train a machine learning model. In this method, humans or existing systems provide correct labels for the training examples, and the model learns the relationship between input data and the expected output labels. After training, the machine can apply what it has learned to correctly classify or label current and future data points.

Unsupervised learning is incorrect because it works with unlabeled data and discovers hidden patterns, groups, or structures without human-provided labels. Reinforcement learning is also incorrect because it is based on actions, rewards, penalties, and learning through interaction with an environment. Semi-supervised learning uses a combination of a small amount of labeled data and a larger amount of unlabeled data, but the question clearly states that humans are labeling the data. “Semi Reinforcement learning” is not the standard answer here. Therefore, the correct choice is A. supervised learning .

Solution architecture provides the structured blueprint needed to move from a business or technical concept to a working implementation. It defines how different systems, applications, data flows, technologies, security requirements, and business needs will fit together. Therefore, it gives teams a solid foundation for developing enterprise software solutions.

A well-defined solution architecture is also valuable when projects become large, complex, or distributed across multiple teams and locations. It creates a common understanding of design decisions, integration points, responsibilities, and technical standards, which supports collaboration and long-term sustainability. In addition, solution architecture helps ensure that the final solution meets business expectations, technical requirements, quality standards, scalability needs, security controls, and operational goals.

Since options A, B, and C all describe valid benefits of solution architecture, the most complete and correct answer is E. All of the above .

The correct answer is A. Weak AI . Artificial Narrow Intelligence, or ANI, is commonly called Weak AI because it is designed to perform a specific task or a limited set of tasks within a defined domain. Examples include recommendation engines, search engines, spam filters, facial recognition systems, voice assistants, fraud detection tools, and chatbots. These systems can perform their assigned functions effectively, but they do not possess general intelligence, consciousness, self-awareness, or human-like understanding across all domains.

Strong AI and General AI refer to Artificial General Intelligence, which would be capable of broad reasoning, learning, and problem-solving across many tasks like a human. SuperAI refers to a theoretical level of intelligence beyond human capability. ExpertAI is not the standard expression for ANI. Since ANI is task-specific and limited in scope, it is correctly expressed as Weak AI .

The correct answer is E. All of the above because the Value Engineering Framework focuses on identifying, delivering, and expanding measurable business value from data and AI initiatives. “Define value creation” is a key step because organizations must first clarify the business problem, expected outcomes, success metrics, stakeholders, and value drivers before investing in an AI solution.

“Realize value creation” is also correct because value must be converted from a planned objective into actual operational or financial impact. This may involve deploying the solution, measuring results, improving processes, reducing cost, increasing revenue, improving risk management, or enhancing customer outcomes.

“Scale value creation” is correct because successful AI initiatives should not remain limited to isolated pilots. Organizations need to scale proven use cases across teams, business units, workflows, and enterprise platforms to maximize return on investment and long-term impact. Since all three options represent steps in value engineering, the best answer is E. All of the above .

The correct answer is D. Restricted Learning because it is not commonly recognized as a standard type of machine learning. The main learning approaches include supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, and transfer learning. Supervised learning uses labeled datasets to train models for prediction or classification. Unsupervised learning uses unlabeled data to discover patterns, clusters, or hidden structures. Semi-supervised learning combines a small amount of labeled data with a larger amount of unlabeled data. Transfer learning reuses knowledge from a pre-trained model and adapts it to a new related task.

“Restricted Learning” is not a standard machine learning category in this context. Although some specific technical terms may include the word “restricted,” such as restricted Boltzmann machines, that does not make “restricted learning” a recognized general type of machine learning. Therefore, the option that is NOT a type of machine learning is D. Restricted Learning .

The correct answer is A. Machine Learning . Deep learning is a specialized subset of machine learning that uses artificial neural networks with multiple layers to learn patterns from data. These layered neural networks can automatically discover features and representations from large datasets, which makes deep learning especially useful for image recognition, speech recognition, natural language processing, recommendation systems, and generative AI applications.

Artificial intelligence is the broader field that includes machine learning, expert systems, reasoning systems, robotics, natural language processing, and other intelligent technologies. Machine learning is a branch within artificial intelligence, and deep learning is a further subset within machine learning. Artificial Narrow Intelligence refers to AI systems designed for specific tasks, while Artificial General Intelligence refers to a theoretical system with broad human-like intelligence. Since deep learning is most directly and correctly classified as a subset of machine learning, the best answer is A .

The correct answer is D. a and b only because statements A and B are the most incorrect for retrospective/concurrent risk-scoring methodology. Retrospective/concurrent risk assessment is mainly used to evaluate model risk based on past or present evidence, current model behavior, observed incidents, model performance changes, risk indicators, and investigation findings. It is not primarily a future-prediction method.

Statement A is incorrect because it says retrospective/concurrent methods “predict” model risk after analyzing historical model performance. Historical performance may be reviewed, but retrospective/concurrent risk scoring is more about assessing or investigating past and current risk conditions, not predicting future risk. Statement B is also incorrect because using current model risk to predict overall model risk for future cycles describes prospective risk, not retrospective/concurrent risk. Statement C is correct because retrospective/concurrent review is suitable when there are changes in model behavior, risk indicators, attacks, data loss, or investigation needs. Therefore, the most incorrect statements are A and B only .