Quiz: AI and Machine Learning Fundamentals — 21 questions

1. Which network task is a direct application of predictive AI?

A. Generating Python scripts for device configuration

B. Forecasting traffic patterns from network data

C. Designing a new network topology

D. Writing documentation for network devices

Answer: (B)

Explanation

Predictive AI uses data to forecast network conditions such as traffic patterns. Generating scripts, designs, and documentation are examples of artifacts produced by generative AI.

2. A system receives customer records without category labels and discovers groups of customers with similar behavior. Which learning type is being used?

A. Unsupervised learning, because the system discovers groupings from unlabeled data

B. Reinforcement learning, because the system receives penalties for incorrect group assignments

C. Deep learning, because every grouping task requires a multilayered neural network

D. Supervised learning, because the system predicts categories supplied during training

Answer: (A)

Explanation

Unsupervised learning finds patterns, relationships, or groupings in data without predefined labels, matching the customer-segmentation situation. Supervised learning requires provided answers, while the scenario does not involve rewards or require a particular neural-network architecture.

3. What distinguishes deep learning from machine learning more generally?

A. Deep learning learns through rewards while other machine learning avoids environmental interaction

B. Deep learning uses multilayered neural networks for complex tasks and large datasets

C. Deep learning trains from labeled answers while other machine learning uses no data labels

D. Deep learning searches for groupings in unlabeled data without making predictions

Answer: (B)

Explanation

Deep learning is a specialized subset of machine learning that uses multilayered neural networks to process large datasets and perform complex tasks. Labels, rewards, and unlabeled grouping describe particular learning setups, not the defining distinction of deep learning.

4. Which combination most directly explains the growing importance of artificial intelligence?

A. Greater computing power, abundant big data, and advances in AI research

B. Expanded software licensing, standardized interfaces, and declining data collection

C. Faster internet access, cheaper hardware, and reduced use of digital services

D. More manual data processing, smaller datasets, and fewer research breakthroughs

Answer: (A)

Explanation

AI has become more important because computing power has increased, large datasets have become available, and AI research has made significant breakthroughs. Faster internet or software licensing may support technology generally but does not capture the stated drivers of AI’s growth.

5. A retailer wants to discover previously unknown customer segments without assigning categories in advance. Which approach best fits this goal?

A. Unsupervised learning, because it clusters data by shared features

B. Reinforcement learning, because it rewards profitable purchases

C. Generative learning, because it produces synthetic customer profiles

D. Supervised learning, because it requires predefined customer labels

Answer: (A)

Explanation

Unsupervised learning can group similar data points into clusters and reveal hidden patterns without predefined labels. Supervised learning would require labeled examples representing the customer categories in advance.

6. How does reinforcement learning improve a model’s behavior over time?

A. It interacts with an environment and uses rewards or penalties to guide actions

B. It studies labeled examples and copies the correct answer for each new input

C. It processes large datasets through several layers of a neural network

D. It groups unlabeled records according to similarities among their features

Answer: (A)

Explanation

Reinforcement learning improves behavior through interaction with an environment, using rewards and penalties to encourage actions that maximize performance over time. Labeled examples describe supervised learning, grouping describes unsupervised learning, and multilayered networks describe deep learning.

7. When a network outage occurs, which Cisco Catalyst Center capability analyzes its underlying cause and may recommend a resolution?

A. AI Network Analytics

B. AI-enhanced Radio Resource Management

C. AI Endpoint Analytics

D. Machine Reasoning Engine

Answer: (D)

Explanation

The Machine Reasoning Engine uses AI to perform root-cause analysis and can suggest or carry out corrective actions. AI Endpoint Analytics instead focuses on identifying and classifying devices.

8. What feedback mechanism allows a reinforcement-learning agent to improve its behavior?

A. It receives a reward or penalty after taking an action

B. It compares its output with a biological neural network

C. It extracts increasingly abstract features through hidden layers

D. It receives a label for every training example before acting

Answer: (A)

Explanation

A reinforcement-learning agent interacts with an environment, takes actions, and uses resulting rewards or penalties to learn which actions produce better outcomes. Receiving labels for training examples is characteristic of supervised learning.

9. Which sequence describes how data moves through a deep-learning network?

A. Input layer, multiple hidden layers, then output layer

B. Hidden layer, labeled examples, then biological processing layer

C. Output layer, one hidden layer, then input layer

D. Input layer, reward layer, then penalty and output layers

Answer: (A)

Explanation

Deep-learning networks pass data from an input layer through multiple hidden layers and finally to an output layer, with successive layers extracting more abstract features. Rewards and penalties belong to reinforcement learning rather than the standard layer sequence.

10. A security team needs to identify an unauthorized device and place it into an appropriate network segment; which capability best fits this task?

A. AI-enhanced Radio Resource Management

B. Machine Reasoning Engine

C. AI Network Analytics

D. AI Endpoint Analytics

Answer: (D)

Explanation

AI Endpoint Analytics identifies and classifies devices, detects unauthorized or unusual behavior, and supports automated profiling and segmentation. The Machine Reasoning Engine focuses on diagnosing the causes of network problems.

11. What does artificial intelligence enable computers to do?

A. Simulate human-like abilities such as recognizing patterns and solving problems

B. Execute fixed instructions without adapting to changing information

C. Store large datasets without interpreting or acting on their contents

D. Connect devices through networks without making decisions from data

Answer: (A)

Explanation

Artificial intelligence enables computers to simulate abilities commonly associated with human intelligence, including pattern recognition, learning, decision-making, and problem-solving. The alternatives describe storage, fixed programming, or networking rather than intelligent behavior.

12. How does supervised learning classify an unseen data point?

A. It groups the point with similar examples without using labels

B. It learns relationships between labeled inputs and their known outputs

C. It generates a new example by recombining learned patterns

D. It selects an action by maximizing rewards from an environment

Answer: (B)

Explanation

Supervised learning studies labeled examples to learn the relationship between each input and its label, then applies that relationship to unseen data. Grouping similar points without labels describes unsupervised learning, not supervised classification.

13. What is the primary role of AI Network Analytics in Cisco Catalyst Center?

A. Establishing behavioral baselines and detecting network anomalies

B. Identifying unauthorized devices and automating device segmentation

C. Performing root-cause analysis and applying corrective actions

D. Adjusting wireless radio settings to balance access-point load

Answer: (A)

Explanation

AI Network Analytics learns normal network behavior, offers optimization recommendations, and monitors for anomalies. Root-cause analysis belongs to the Machine Reasoning Engine, not AI Network Analytics.

14. What is an artificial neural network?

A. A computational model inspired by biological information processing

B. A reward-based system that learns actions through environmental feedback

C. A living information-processing system found in the human brain

D. A clustering method that groups data without human supervision

Answer: (A)

Explanation

An artificial neural network is a computational model inspired by the way biological neural networks process information. A biological neural network is a living system, whereas an artificial one is implemented computationally.

15. How does a machine-learning algorithm typically make a prediction about new data?

A. It identifies patterns in input data and applies them to make a prediction or decision

B. It removes unusual observations before a human chooses the final answer

C. It follows a complete set of manually written instructions for every possible input

D. It copies the most recent example without analyzing relationships among the data

Answer: (A)

Explanation

Machine-learning algorithms detect patterns in input data and use those patterns to make predictions or decisions on new data. A hard-coded system instead relies on explicit human-written instructions, which is the key distinction highlighted here.

16. What kind of data is used to train a supervised-learning model?

A. Labeled data containing the correct answers for the training examples

B. Unlabeled data from which the model must discover its own groupings

C. Large datasets processed through multilayered neural networks

D. Interaction data paired with rewards or penalties for chosen actions

Answer: (A)

Explanation

Supervised learning trains on labeled examples for which the correct answers are provided, allowing the model to learn predictions or classifications. Unlabeled data characterizes unsupervised learning, while rewards and multilayered networks describe other learning types.

17. A network engineer asks an AI system to produce a Python script that configures multiple routers; which type of AI is being used?

A. Predictive AI

B. Radio resource management AI

C. Endpoint classification AI

D. Generative AI

Answer: (D)

Explanation

Generative AI creates network artifacts, including Python automation scripts for configuring devices. Predictive AI instead forecasts network behavior, threats, or equipment failures.

18. Which statement best defines machine learning in relation to artificial intelligence?

A. It is a separate field that replaces AI by following instructions written in advance

B. It is a branch of robotics that lets machines move without processing information

C. It is a data-storage method that organizes examples for later human analysis

D. It is a subset of AI in which computers learn from data and improve without explicit programming

Answer: (D)

Explanation

Machine learning is a subset of artificial intelligence that allows systems to learn from data and improve without explicit programming for every behavior. The broader AI field can also include applications that do not learn from data, so ML does not replace AI.

19. A hospital uses past patient records to estimate the likelihood of future treatment outcomes. Which type of AI is being used?

A. Reinforcement AI, because it learns treatment actions from penalties

B. Unsupervised AI, because it groups patients without examining outcomes

C. Predictive AI, because it forecasts outcomes from historical data

D. Generative AI, because it creates new medical content from patterns

Answer: (C)

Explanation

Predictive AI analyzes historical data to forecast future outcomes, including healthcare outcomes. Generative AI instead learns patterns to create new content such as text, images, or audio.

20. How does AI-enhanced Radio Resource Management improve wireless network operation?

A. It analyzes incidents to determine causes and recommend corrective actions.

B. It classifies endpoints and segments devices according to their identities.

C. It establishes behavioral baselines to predict anomalies across the network.

D. It dynamically adjusts radio settings to balance load and reduce interference.

Answer: (D)

Explanation

AI-enhanced Radio Resource Management dynamically changes radio settings to balance wireless demand, limit interference, and improve coverage. Establishing baselines is handled by AI Network Analytics, while incident diagnosis is handled by the Machine Reasoning Engine.

21. Which task is a direct example of generative AI?

A. Forecasting tomorrow’s weather from historical measurements

B. Detecting an unusual pattern in network activity

C. Creating an image from learned visual patterns

D. Predicting traffic congestion on a planned route

Answer: (C)

Explanation

Generative AI learns patterns from existing data and creates new content, such as an image. Weather forecasting, traffic prediction, and anomaly detection are applications of predictive AI.