CBSE Class 10 AI Unit 2 MCQs with Answers & Explanations (2026–27) | Advanced Concepts of Modeling in AI
Class 10 · Artificial Intelligence
Prepare for your CBSE Class 10 Artificial Intelligence (Code 417) Examination with these 50 important Multiple Choice Questions (MCQs) from Unit 2: Advanced Concepts of Modeling in AI. Every question includes the correct answer along with a detailed explanation to help you understand key AI concepts and perform well in CBSE examinations.
Topics Covered
- Artificial Intelligence, Machine Learning & Deep Learning
- Features, Labels & Datasets
- Rule-Based and Learning-Based Approaches
- Training & Testing Datasets
- Supervised Learning
- Unsupervised Learning
- Reinforcement Learning
- Artificial Neural Networks (ANN)
- Convolutional Neural Networks (CNN)
- Perceptron, Weights, Bias & Threshold
- Competency-Based MCQs
CBSE Exam Tip: Understand the differences between Machine Learning, Deep Learning, Rule-Based Systems, Supervised, Unsupervised, and Reinforcement Learning. Practice identifying the correct learning model for real-life examples, as these are commonly tested in competency-based and case-based MCQs.
Multiple Choice Questions (MCQs)
Q1. Which of the following terms acts as the "umbrella" terminology covering both Machine Learning and Deep Learning?
A. Data Science
B. Computer Vision
C. Artificial Intelligence
D. Natural Language Processing
Show Answer & Explanation
Correct Answer: C. Artificial Intelligence
Explanation: Artificial Intelligence (AI) is the broadest field of study. Machine Learning (ML) is a subset of AI, and Deep Learning (DL) is a subset of Machine Learning. This relationship is often represented using the Funnel Approach.
Why other options are incorrect:
- A. Data Science focuses on extracting insights from data but is not the umbrella term.
- B. Computer Vision is an AI domain that enables machines to understand images.
- D. Natural Language Processing (NLP) is another AI domain that deals with human language.
Q2. In a dataset, what are the columns that describe different properties of the data called?
A. Labels
B. Features
C. Rows
D. Tokens
Show Answer & Explanation
Correct Answer: B. Features
Explanation: Features are the characteristics or properties of data. For example, in a fruit dataset, Color, Weight, and Size are features that help identify the fruit.
Why other options are incorrect:
- A. Labels represent the output or target value.
- C. Rows represent individual records in the dataset.
- D. Tokens are mainly used in Natural Language Processing.
Q3. What is the primary characteristic of Deep Learning?
A. It strictly follows rules defined by the programmer.
B. It enables software to learn automatically using large amounts of data.
C. It works only with small datasets.
D. It is the simplest form of Artificial Intelligence.
Show Answer & Explanation
Correct Answer: B. It enables software to learn automatically using large amounts of data.
Explanation: Deep Learning is the most advanced form of Machine Learning. It uses Artificial Neural Networks with multiple hidden layers to learn complex patterns from huge amounts of data.
Why other options are incorrect:
- A. Rule-Based Systems follow predefined rules, not Deep Learning.
- C. Deep Learning performs best with large datasets.
- D. Deep Learning is the most advanced, not the simplest, AI technique.
Q4. Which type of data contains an attached tag or identification such as a name, category, or number?
A. Unlabeled Data
B. Raw Data
C. Discrete Data
D. Labeled Data
Show Answer & Explanation
Correct Answer: D. Labeled Data
Explanation: Labeled data contains predefined tags or categories that help AI models learn the relationship between input data and the expected output during training.
Why other options are incorrect:
- A. Unlabeled data has no predefined tags.
- B. Raw data may or may not have labels.
- C. Discrete data refers to categorical values, not necessarily labeled data.
Q5. A model that follows instructions written by the developer and cannot improve by learning from new data uses which approach?
A. Learning-Based Approach
B. Rule-Based Approach
C. Reinforcement Learning
D. Unsupervised Learning
Show Answer & Explanation
Correct Answer: B. Rule-Based Approach
Explanation: In a Rule-Based Approach, the developer writes all the decision rules. The model cannot modify these rules automatically or learn from experience.
Why other options are incorrect:
- A. Learning-Based models improve automatically using data.
- C. Reinforcement Learning learns through rewards and penalties.
- D. Unsupervised Learning discovers patterns from unlabeled data.
Q6. Which dataset is used to evaluate the performance of an AI model after training?
A. Training Dataset
B. Feature Dataset
C. Testing Dataset
D. Label Dataset
Show Answer & Explanation
Correct Answer: C. Testing Dataset
Explanation: The Testing Dataset contains unseen data that is used to check how accurately the trained AI model performs on new examples.
Why other options are incorrect:
- A. The Training Dataset is used to teach the model.
- B. "Feature Dataset" is not a standard dataset type.
- D. Labels are part of data, not a separate dataset.
Q7. In a Learning-Based Approach, how does the computer learn to solve problems?
A. By following fixed rules forever.
B. By analyzing examples and learning from feedback.
C. By randomly guessing every answer.
D. By ignoring historical data.
Show Answer & Explanation
Correct Answer: B. By analyzing examples and learning from feedback.
Explanation: Learning-Based AI improves by studying data, identifying patterns, and learning from previous experiences or feedback instead of relying on fixed rules.
Why other options are incorrect:
- A. This describes Rule-Based AI.
- C. AI learning is based on data, not random guessing.
- D. Historical data plays an important role in learning.
Q8. Which Machine Learning category uses labeled examples and a "teacher" during training?
A. Unsupervised Learning
B. Reinforcement Learning
C. Deep Learning
D. Supervised Learning
Show Answer & Explanation
Correct Answer: D. Supervised Learning
Explanation: Supervised Learning uses labeled datasets where the correct answers are already known. The model learns by comparing its predictions with these correct answers.
Why other options are incorrect:
- A. Unsupervised Learning uses unlabeled data.
- B. Reinforcement Learning learns through rewards.
- C. Deep Learning is a technique, not a learning category.
Q9. Predicting whether a customer is eligible for a bank loan (Yes/No) is an example of:
A. Regression Model
B. Clustering Model
C. Classification Model
D. Association Model
Show Answer & Explanation
Correct Answer: C. Classification Model
Explanation: Classification models predict discrete categories such as Yes/No, Pass/Fail, Approved/Rejected, or Spam/Not Spam.
Why other options are incorrect:
- A. Regression predicts continuous values.
- B. Clustering groups similar data.
- D. Association discovers relationships between items.
Q10. Which learning approach discovers hidden patterns and relationships from unlabeled data without any teacher?
A. Supervised Learning
B. Unsupervised Learning
C. Rule-Based Modeling
D. Neural Network Training
Show Answer & Explanation
Correct Answer: B. Unsupervised Learning
Explanation: Unsupervised Learning analyzes unlabeled data to identify natural groupings, trends, and hidden relationships without any predefined answers.
Why other options are incorrect:
- A. Supervised Learning requires labeled data.
- C. Rule-Based Modeling follows programmer-defined rules.
- D. Neural Networks can be used in different learning approaches but are not themselves a learning approach.
Q11. A model used to predict a continuous value such as house prices or temperature is called a:
A. Classification Model
B. Clustering Model
C. Regression Model
D. Rule-Based Model
Show Answer & Explanation
Correct Answer: C. Regression Model
Explanation: A Regression Model predicts continuous numerical values such as house prices, temperature, salary, rainfall, or stock prices based on input features.
Why other options are incorrect:
- A. Classification predicts categories like Yes/No or Pass/Fail.
- B. Clustering groups similar data points.
- D. Rule-Based Models follow predefined rules instead of learning patterns.
Q12. Which Unsupervised Learning technique finds relationships between items, such as products frequently purchased together?
A. Clustering
B. Regression
C. Classification
D. Association
Show Answer & Explanation
Correct Answer: D. Association
Explanation: Association Learning discovers interesting relationships between items. For example, supermarkets use it to identify products that customers often buy together.
Why other options are incorrect:
- A. Clustering creates groups of similar data.
- B. Regression predicts continuous values.
- C. Classification predicts predefined categories.
Q13. In Reinforcement Learning, how does the machine improve its performance?
A. By following predefined rules only
B. By learning through trial-and-error and rewards
C. By using only labeled datasets
D. By grouping similar data into clusters
Show Answer & Explanation
Correct Answer: B. By learning through trial-and-error and rewards
Explanation: Reinforcement Learning allows an AI agent to interact with its environment, receive rewards or penalties, and gradually learn the best actions through repeated practice.
Why other options are incorrect:
- A. Rule-Based systems use predefined rules.
- C. Labeled datasets are used in Supervised Learning.
- D. Clustering belongs to Unsupervised Learning.
Q14. Which of the following is a real-life application of Reinforcement Learning?
A. Email spam detection
B. Predicting house prices
C. A humanoid robot learning to walk
D. Classifying fruits by color
Show Answer & Explanation
Correct Answer: C. A humanoid robot learning to walk
Explanation: A humanoid robot learns by repeatedly trying different movements, receiving rewards for successful actions, and improving its performance over time.
Why other options are incorrect:
- A. Spam detection commonly uses Supervised Learning.
- B. House price prediction uses Regression.
- C. Fruit classification is a Classification task.
Q15. What is the major advantage of Artificial Neural Networks (ANN)?
A. They require no data.
B. They automatically extract useful features from data.
C. They only work with rule-based systems.
D. They are limited to mathematical calculations.
Show Answer & Explanation
Correct Answer: B. They automatically extract useful features from data.
Explanation: Artificial Neural Networks automatically identify important patterns and features from large datasets without requiring the programmer to manually define every rule.
Why other options are incorrect:
- A. Neural Networks require large amounts of data.
- C. They learn from data instead of relying only on fixed rules.
- D. They are widely used for image recognition, speech recognition, NLP, and many AI applications.
Q16. Which layer of an Artificial Neural Network receives input data from the user?
A. Hidden Layer
B. Output Layer
C. Processing Layer
D. Input Layer
Show Answer & Explanation
Correct Answer: D. Input Layer
Explanation: The Input Layer receives the raw input data and passes it to the hidden layers for processing.
Why other options are incorrect:
- A. Hidden Layers perform computations.
- B. The Output Layer produces the final result.
- C. "Processing Layer" is not the standard term used in ANN.
Q17. In an Artificial Neural Network, where does the actual processing of information take place?
A. Input Layer
B. Hidden Layers
C. Output Layer
D. Dataset Layer
Show Answer & Explanation
Correct Answer: B. Hidden Layers
Explanation: Hidden Layers perform mathematical computations using weights and biases to identify patterns and relationships in the data.
Why other options are incorrect:
- A. The Input Layer only accepts data.
- C. The Output Layer only displays the final prediction.
- D. Dataset Layer is not a valid ANN layer.
Q18. In the Perceptron model, what do "Weights" represent?
A. The total number of neurons
B. The importance assigned to each input feature
C. The output produced by the model
D. The training dataset
Show Answer & Explanation
Correct Answer: B. The importance assigned to each input feature
Explanation: Weights determine how much influence each input has on the final decision. Larger weights indicate more important features.
Why other options are incorrect:
- A. Weights do not count neurons.
- C. They influence the output but are not the output.
- D. Weights are mathematical values, not datasets.
Q19. What is the function of "Bias" in a Perceptron?
A. It acts as the input feature.
B. It adjusts the decision-making behavior of the model.
C. It stores the training data.
D. It replaces the weights.
Show Answer & Explanation
Correct Answer: B. It adjusts the decision-making behavior of the model.
Explanation: Bias shifts the decision boundary, allowing the model to become more flexible and make better predictions even when input values are zero.
Why other options are incorrect:
- A. Bias is not an input feature.
- C. Training data is stored separately.
- D. Bias works together with weights; it does not replace them.
Q20. Which Deep Learning algorithm is specially designed to recognize and classify images?
A. Artificial Neural Network (ANN)
B. Logistic Regression
C. Decision Tree
D. Convolutional Neural Network (CNN)
Show Answer & Explanation
Correct Answer: D. Convolutional Neural Network (CNN)
Explanation: CNNs are specialized Deep Learning models designed for image processing. They automatically identify important features such as edges, shapes, and objects within images.
Why other options are incorrect:
- A. ANN is a general Neural Network architecture.
- B. Logistic Regression is mainly used for classification.
- C. Decision Trees are Machine Learning algorithms but are not specialized for image recognition.
Q21. Artificial Neural Networks (ANNs) are loosely modeled after which biological system?
A. Human digestive system
B. Neurons in the human brain
C. Human circulatory system
D. Plant cells
Show Answer & Explanation
Correct Answer: B. Neurons in the human brain
Explanation: Artificial Neural Networks are inspired by the structure and functioning of biological neurons in the human brain. Just as neurons communicate with each other, artificial neurons process and transmit information through interconnected layers.
Why other options are incorrect:
- A. The digestive system has no relation to AI learning models.
- C. The circulatory system transports blood, not information.
- D. Plant cells are unrelated to Neural Networks.
Q22. Which layer of an Artificial Neural Network provides the final prediction or output?
A. Input Layer
B. Hidden Layer
C. Output Layer
D. Training Layer
Show Answer & Explanation
Correct Answer: C. Output Layer
Explanation: The Output Layer receives the processed information from the hidden layers and presents the final prediction or decision to the user.
Why other options are incorrect:
- A. The Input Layer only receives data.
- B. Hidden Layers perform internal computations.
- D. "Training Layer" is not a valid ANN layer.
Q23. During the learning process, what happens to the weights in a Neural Network?
A. They remain fixed forever.
B. They are removed after every calculation.
C. They are adjusted based on the prediction error.
D. They are selected randomly every time.
Show Answer & Explanation
Correct Answer: C. They are adjusted based on the prediction error.
Explanation: During training, the Neural Network compares its prediction with the correct answer. The weights are continuously adjusted to reduce errors and improve prediction accuracy.
Why other options are incorrect:
- A. Fixed weights would prevent learning.
- B. Weights are updated, not deleted.
- D. Weight adjustment follows mathematical optimization rather than random selection.
Q24. In the "Go to the Park" Perceptron example, if the calculated output is greater than the threshold value, what will be the decision?
A. No
B. Yes
C. Retry
D. Error
Show Answer & Explanation
Correct Answer: B. Yes
Explanation: In a Perceptron, if the final calculated value is greater than the threshold, the output is considered positive, resulting in a "Yes" decision.
Why other options are incorrect:
- A. "No" is selected only when the output is below the threshold.
- C. No retry is required.
- D. This situation does not indicate an error.
Q25. Which term refers to information in any form, such as numbers, text, images, or tables?
A. Feature
B. Label
C. Data
D. Node
Show Answer & Explanation
Correct Answer: C. Data
Explanation: Data is raw information that can exist in various forms, including numbers, text, images, videos, and tables. AI systems learn from this data.
Why other options are incorrect:
- A. Features are properties of the data.
- B. Labels represent the expected output.
- C. Nodes are computational units in Neural Networks.
Q26. Which statement is NOT true about Rule-Based AI models?
A. They are adaptive to changing data.
B. They are static in nature.
C. They follow developer-defined rules.
D. They may fail when new situations arise.
Show Answer & Explanation
Correct Answer: A. They are adaptive to changing data.
Explanation: Rule-Based AI systems do not learn automatically. They follow predefined rules and cannot adapt to changing data without human intervention.
Why other options are incorrect:
- B. Rule-Based models are static.
- C. They rely completely on developer-written rules.
- D. They often fail when faced with situations not covered by the rules.
Q27. What is Clustering in Unsupervised Learning?
A. Predicting continuous values
B. Dividing similar data into groups or clusters
C. Assigning labels to data
D. Following fixed rules
Show Answer & Explanation
Correct Answer: B. Dividing similar data into groups or clusters
Explanation: Clustering groups data points with similar characteristics without using predefined labels. It is one of the most common techniques in Unsupervised Learning.
Why other options are incorrect:
- A. Regression predicts continuous values.
- C. Labels are used in Supervised Learning.
- D. Fixed rules belong to Rule-Based AI.
Q28. In Supervised Learning, if "Color" is used to predict the "Fruit Name", then "Color" is called:
A. Label
B. Output
C. Feature
D. Target
Show Answer & Explanation
Correct Answer: C. Feature
Explanation: Features are the input variables used by the AI model to make predictions. In this example, "Color" is the input feature used to predict the fruit name.
Why other options are incorrect:
- A. Fruit Name is the label.
- B. Output is the prediction made by the model.
- D. Target refers to the expected output, not the input.
Q29. A music recommendation system that groups songs based on tempo, mood, or intensity mainly uses:
A. Supervised Learning
B. Reinforcement Learning
C. Rule-Based Approach
D. Unsupervised Learning
Show Answer & Explanation
Correct Answer: D. Unsupervised Learning
Explanation: Music recommendation systems often use clustering to group songs with similar characteristics and recommend similar songs to users.
Why other options are incorrect:
- A. Supervised Learning requires labeled data.
- B. Reinforcement Learning learns through rewards.
- C. Rule-Based systems cannot automatically discover song similarities.
Q30. In the Apple vs. Cherry Reinforcement Learning example, what happens when the machine receives negative feedback?
A. It stops learning permanently.
B. It ignores the feedback.
C. It learns from its mistake and improves future predictions.
D. It deletes the training dataset.
Show Answer & Explanation
Correct Answer: C. It learns from its mistake and improves future predictions.
Explanation: Negative feedback acts as a penalty in Reinforcement Learning. The AI agent learns that its previous action was incorrect and modifies its behavior to make better decisions in future attempts.
Why other options are incorrect:
- A. Learning continues after receiving feedback.
- B. Feedback is essential for improving performance.
- D. The dataset is not deleted because of incorrect predictions.
Q31. Which statement best describes Machine Learning?
A. A system that never changes its rules
B. Algorithms that enable machines to learn from data and improve with experience
C. A technique that only processes images
D. A method where every outcome is manually programmed
Show Answer & Explanation
Correct Answer: B. Algorithms that enable machines to learn from data and improve with experience
Explanation: Machine Learning enables computers to learn patterns from historical data, make predictions, and continuously improve their performance without being explicitly programmed for every situation.
Why other options are incorrect:
- A. This describes a Rule-Based Approach.
- C. Machine Learning is not limited to image processing.
- D. Manual programming is a Rule-Based system, not Machine Learning.
Q32. A system groups all cats into one cluster and all dogs into another without being told what a "cat" or "dog" is. This is an example of:
A. Classification
B. Regression
C. Clustering
D. Association
Show Answer & Explanation
Correct Answer: C. Clustering
Explanation: Clustering is an Unsupervised Learning technique where similar data points are automatically grouped together without predefined labels.
Why other options are incorrect:
- A. Classification requires labeled data.
- B. Regression predicts numerical values.
- D. Association discovers relationships between items.
Q33. Which term refers to special identifiers attached to data to provide meaning?
A. Weights
B. Biases
C. Features
D. Labels
Show Answer & Explanation
Correct Answer: D. Labels
Explanation: Labels provide the expected output or category for each data record. They help supervised learning algorithms understand the correct answer during training.
Why other options are incorrect:
- A. Weights determine the importance of features.
- B. Bias adjusts the decision boundary.
- C. Features are input characteristics, not output identifiers.
Q34. If an AI learns to play a video game by earning points for correct actions, which learning approach is being used?
A. Supervised Learning
B. Reinforcement Learning
C. Unsupervised Learning
D. Rule-Based Approach
Show Answer & Explanation
Correct Answer: B. Reinforcement Learning
Explanation: Reinforcement Learning uses rewards and penalties. The AI agent learns by interacting with the environment and maximizing the total reward.
Why other options are incorrect:
- A. Supervised Learning requires labeled examples.
- C. Unsupervised Learning does not use rewards.
- D. Rule-Based systems do not learn from experience.
Q35. The diagram showing Artificial Intelligence as the largest circle, Machine Learning inside it, and Deep Learning inside Machine Learning is known as the:
A. AI Project Cycle
B. Rule-Based Model
C. Funnel Approach
D. Neural Network Model
Show Answer & Explanation
Correct Answer: C. Funnel Approach
Explanation: The Funnel Approach illustrates the relationship among AI, Machine Learning, and Deep Learning, where each inner technology is a subset of the outer one.
Why other options are incorrect:
- A. The AI Project Cycle describes AI development stages.
- B. Rule-Based Models are a different AI approach.
- D. Neural Networks are Deep Learning models.
Q36. Which of the following is an example of discrete data?
A. Temperature in Jaipur
B. Height of students
C. Student Grades (A, B, C, D)
D. Weight of fruits
Show Answer & Explanation
Correct Answer: C. Student Grades (A, B, C, D)
Explanation: Discrete data consists of distinct categories or classes. Student grades are fixed categories rather than continuous numerical values.
Why other options are incorrect:
- A. Temperature is continuous data.
- B. Height is continuous data.
- D. Weight is measured on a continuous scale.
Q37. What is the primary function of the Output Layer in a Neural Network?
A. Receive raw input data
B. Perform feature extraction
C. Provide the final prediction or result
D. Adjust weights automatically
Show Answer & Explanation
Correct Answer: C. Provide the final prediction or result
Explanation: The Output Layer presents the final prediction after all computations have been completed in the hidden layers.
Why other options are incorrect:
- A. Raw input is received by the Input Layer.
- B. Hidden Layers perform feature extraction.
- D. Weights are updated during training, not by the Output Layer.
Q38. In the Perceptron model, the weighted sum of inputs is compared with a:
A. Bias
B. Cluster
C. Label
D. Threshold
Show Answer & Explanation
Correct Answer: D. Threshold
Explanation: The Perceptron compares the weighted sum of inputs with a threshold value. If the sum exceeds the threshold, the output is generally "Yes" or 1; otherwise, it is "No" or 0.
Why other options are incorrect:
- A. Bias shifts the decision boundary but is not the comparison value.
- B. Clusters belong to Unsupervised Learning.
- C. Labels represent expected outputs.
Q39. Unlabeled data is best described as:
A. Data containing predefined categories
B. Data with solved answers
C. Raw data without tags or labels
D. Data used only in Rule-Based systems
Show Answer & Explanation
Correct Answer: C. Raw data without tags or labels
Explanation: Unlabeled data contains no predefined output or category. AI models must identify patterns or relationships on their own.
Why other options are incorrect:
- A. This describes labeled data.
- B. Solved answers are used in Supervised Learning.
- D. Rule-Based systems do not depend on unlabeled data.
Q40. Regression algorithms are mainly used for:
A. Predicting continuous numerical values
B. Grouping similar customers
C. Categorizing emails as spam or not spam
D. Detecting faces in images
Show Answer & Explanation
Correct Answer: A. Predicting continuous numerical values
Explanation: Regression models predict values on a continuous scale, such as salary, rainfall, temperature, or house prices.
Why other options are incorrect:
- B. Customer grouping uses Clustering.
- C. Spam detection is a Classification problem.
- D. Face detection is generally performed using Computer Vision techniques.
Q41. Artificial Neural Networks (ANNs) are most useful when:
A. The rules are very simple.
B. The dataset is very large.
C. No data is available.
D. Every step must be manually guided by a human.
Show Answer & Explanation
Correct Answer: B. The dataset is very large.
Explanation: Artificial Neural Networks perform exceptionally well with large datasets. They automatically learn complex patterns from huge amounts of data, making them suitable for image recognition, speech recognition, and language processing.
Why other options are incorrect:
- A. Simple rule-based problems do not require Neural Networks.
- C. Neural Networks require sufficient training data.
- D. ANNs learn automatically rather than depending on human guidance for every decision.
Q42. An AI model that recognizes a tagged friend in a social media photograph is an example of:
A. Supervised Learning
B. Unsupervised Learning
C. Reinforcement Learning
D. Rule-Based Approach
Show Answer & Explanation
Correct Answer: A. Supervised Learning
Explanation: Tagged photographs provide labeled examples. The AI learns the relationship between facial features and the person's identity using these labels.
Why other options are incorrect:
- B. Unsupervised Learning does not require labels.
- C. Reinforcement Learning learns using rewards and penalties.
- D. Face recognition relies on learned patterns rather than fixed rules.
Q43. Writing algorithms and developing code that enables a machine to behave intelligently is known as:
A. Data Exploration
B. Problem Scoping
C. AI Modelling
D. Evaluation
Show Answer & Explanation
Correct Answer: C. AI Modelling
Explanation: AI Modelling is the process of designing, developing, and training algorithms that allow machines to perform intelligent tasks.
Why other options are incorrect:
- A. Data Exploration analyzes collected data.
- B. Problem Scoping identifies the problem and objectives.
- D. Evaluation measures the model's performance.
Q44. Classification models work mainly with which type of dataset?
A. Continuous Dataset
B. Infinite Dataset
C. Discrete Dataset
D. Sequential Dataset
Show Answer & Explanation
Correct Answer: C. Discrete Dataset
Explanation: Classification models predict categories or classes, making them suitable for discrete datasets such as Pass/Fail, Spam/Not Spam, or Approved/Rejected.
Why other options are incorrect:
- A. Continuous datasets are generally used with Regression models.
- B. Infinite Dataset is not a standard dataset type.
- D. Sequential datasets are used in specialized applications but are not the defining dataset type for classification.
Q45. Which statement correctly describes the Hidden Layer of an Artificial Neural Network?
A. It receives raw input from the user.
B. It performs computations using weights and biases.
C. It displays the final prediction to the user.
D. It stores the testing dataset.
Show Answer & Explanation
Correct Answer: B. It performs computations using weights and biases.
Explanation: Hidden Layers are responsible for processing information. Each neuron performs calculations using weights, biases, and activation functions to identify meaningful patterns.
Why other options are incorrect:
- A. The Input Layer receives raw data.
- C. The Output Layer presents the final result.
- D. Hidden Layers do not store datasets.
Q46. What is a "Node" in an Artificial Neural Network?
A. A pixel in an image
B. A rule written by the programmer
C. A computational unit that performs a specific task within a layer
D. The final prediction generated by the model
Show Answer & Explanation
Correct Answer: C. A computational unit that performs a specific task within a layer.
Explanation: A Node (also called an Artificial Neuron) receives inputs, performs mathematical calculations, applies an activation function, and passes the result to the next layer.
Why other options are incorrect:
- A. Pixels are image elements.
- B. Rules belong to Rule-Based AI.
- D. Predictions are produced by the Output Layer.
Q47. In Unsupervised Learning, what is a Cluster?
A. A predefined label assigned by the programmer
B. A group of data points having similar characteristics
C. A mathematical weight in a Neural Network
D. A type of regression model
Show Answer & Explanation
Correct Answer: B. A group of data points having similar characteristics.
Explanation: A Cluster is formed by grouping data points with similar properties. The AI discovers these similarities automatically without using labeled data.
Why other options are incorrect:
- A. Clusters are formed automatically, not predefined.
- C. Weights belong to Neural Networks.
- D. Regression predicts continuous values.
Q48. Predicting whether a student will "Pass" or "Fail" based on examination marks is an example of:
A. Regression
B. Association
C. Clustering
D. Classification
Show Answer & Explanation
Correct Answer: D. Classification
Explanation: Since the output belongs to one of two categories—Pass or Fail—it is a Classification problem.
Why other options are incorrect:
- A. Regression predicts numerical values.
- B. Association finds relationships between variables.
- C. Clustering groups similar data without labels.
Q49. Which Machine Learning approach uses a reward mechanism to improve performance?
A. Supervised Learning
B. Unsupervised Learning
C. Reinforcement Learning
D. Rule-Based Learning
Show Answer & Explanation
Correct Answer: C. Reinforcement Learning
Explanation: Reinforcement Learning improves performance through rewards and penalties. The AI agent learns the best actions by maximizing rewards over repeated interactions.
Why other options are incorrect:
- A. Supervised Learning uses labeled data.
- B. Unsupervised Learning discovers hidden patterns.
- D. Rule-Based Learning follows predefined instructions without learning from rewards.
Q50. How does a computer represent a grayscale image internally?
A. As a collection of words
B. As a two-dimensional array of pixel values ranging from 0 to 255
C. As three RGB color channels
D. As a set of if-then rules
Show Answer &Explanation
Correct Answer: B. As a two-dimensional array of pixel values ranging from 0 to 255.
Explanation: A grayscale image consists of a single 2D matrix of pixels. Each pixel stores an intensity value from 0 (black) to 255 (white), with intermediate values representing different shades of gray.
Why other options are incorrect:
- A. Images are stored as pixel values, not words.
- C. Three RGB channels are used for color images, not grayscale images.
- D. Images are represented numerically rather than as rules.
? Quick Revision
- Artificial Intelligence (AI) is the umbrella technology that includes Machine Learning (ML) and Deep Learning (DL).
- Features are input variables, while Labels are the expected outputs.
- Rule-Based Systems follow predefined instructions and cannot learn automatically.
- Learning-Based Systems improve by analyzing data and experience.
- Supervised Learning uses labeled data; Unsupervised Learning uses unlabeled data; Reinforcement Learning learns using rewards and penalties.
- Regression predicts continuous values, while Classification predicts categories.
- Artificial Neural Networks (ANNs) are inspired by the human brain and consist of Input, Hidden, and Output Layers.
- Weights represent the importance of input features, while Bias adjusts the decision boundary.
- Convolutional Neural Networks (CNNs) are specialized Deep Learning models for image recognition.
- A grayscale image is stored as a 2D matrix of pixel values ranging from 0 to 255.
? Exam Tip
For the CBSE Class 10 Artificial Intelligence (Code 417) examination, clearly understand the differences between AI, Machine Learning, and Deep Learning; compare Rule-Based and Learning-Based Approaches; remember the three types of Machine Learning; know the difference between Classification and Regression; and revise the structure of Artificial Neural Networks (Input Layer, Hidden Layer, Output Layer). Competency-based and case-based questions are frequently asked from these concepts.