NAVRACHANA HIGHER SECONDARY SCHOOL, SAMA
Here is a comprehensive set of MCQs, short answer, and long answer questions with well- crafted answers, all based
strictly on UNIT-02 , “Advanced Concepts of Modeling in AI”.
Multiple Choice Questions (MCQs) with Answers
Question Options Answer
1. Which approach allows a
B) Learning- based
machine to adapt to changing A) Rule-based B) Learning-based C) Both D) None
data?
2. In which type of learning is A) Supervised B) Unsupervised C) Reinforcement
A) Supervised
data labeled with the desired D) Deep learning
output?
3. What is the primary purpose
A) Assign rewards B) Group similar data C) Predict B) Group similar
of clustering in unsupervised
continuous output D) Label data data
learning?
4. An AI agent learning by trial
A) Supervised learning B) Unsupervised learning C) Reinforcement
and error with rewards and
C) Reinforcement learning D) Regression learning
penalties is using:
A) Requires labeled data B) Used for
5. Which of the following is C) Canʼt handle
classification/regression C) Can’t handle images
NOT true about supervised D) Used in email filters images
learning?
6. Artificial neural networks A) The human brain B) Quantum computers C) A) The human
are inspired by: Mathematical models D) Logics brain
7. The process of adjusting
A) Activation B) Training C) Recognition D)
weights in a neural network B) Training
Clustering
is called:
8. Which model predicts a A) Classification B) Clustering C) Regression D)
C) Regression
continuous value as output? Association
9. Grouping songs based on
A) Classification B) Regression C) Clustering D)
tempo and intensity best C) Clustering
Association
illustrates:
10. The final layer in a neural
network that produces the result A) Input B) Output C) Hidden D) Activation B) Output
for the user is called:
Short Answer Questions with Answers
1. What is the key differencebetween rule-based and learning-based AI models?
Answer:
Rule-based models follow fixed instructions predefined by humans and do not adapt to new data, while
learning-based models learn patterns from data and adapt as new data is introduced, improving their
performance over time.
2. Define supervised,unsupervised,and reinforcement learning with examples.
Answer:
Supervised Learning: Uses labeled data to train models, e.g., classifying emails as spam or not.
Unsupervised Learning: Uses unlabeled data to identify patterns, e.g., clustering customers
by purchasing habits.
Reinforcement Learning: Learns through trial and error using rewards, e.g., teaching a robot to walk by
rewarding successful steps.
3. What is clustering in machine learning? How does it differ from classification?
Answer:
Clustering is grouping data points based on similarities without predefined labels, while classification assigns
data points to specific classes defined prior to training.
4. Explain the main functions of input, hidden, and output layers in a neural network.
Answer:
Input layer: Receives the raw data.
Hidden layer(s): Perform computations, extracting patterns through weighted nodes.
Output layer: Produces the final prediction or result for the user.
5. Differentiate between classification and regression models.
Answer:
Classification models predict discrete classes (e.g., spam/not spam), while regression models predict continuous
values (e.g., house price).
Long Answer Questions with Model Answers
1. Explain the three main categories of learning-based approaches in AI with suitable scenarios.
Answer:
Supervised learning involves training a model on labeled data, where each input has a corresponding output
label. For instance, in email spam detection, the model learns from examples categorized as “spam” or “not
spam.”
Unsupervised learning uses unlabeled datasets; the model identifies patterns or structures within the data. An
example is market segmentation, where customer purchasing patterns are used to group customers without
prior knowledge of segments.
Reinforcement learning is a process in which an agent learns to make decisions by performing actions and
receiving rewards or penalties. For example, a video game AI becomes better at
playing by getting positive feedback for winning moves and negative feedback for losing ones.
2. Describe the structure and functioning of a neural network, highlighting the roles of nodes,
weights, and layers.
Answer:
A neural network consists of interconnected nodes (neurons) organized into layers: an input layer, one or more
hidden layers, and an output layer. Each node receives data, processes it using an assigned weight and a bias,
and passes the result to nodes in the next layer through an activation function.
The input layer captures raw data; hidden layers extract features and perform transformations; the output
layer generates the result. Weights and biases are adjusted during training to minimize errors, enabling the
network to learn complex relationships and representations from large datasets, such as recognizing faces or
predicting prices.
3. Compare and contrast supervised, unsupervised, and reinforcement learning. Discuss their suitability
for real-world applications.
Answer:
Supervised learning is best when abundant labeled data is available and the desired output is clearly defined,
such as in handwriting recognition or disease diagnosis.
Unsupervised learning is suitable when the objective is to explore underlying patterns or groupings in data, like
customer segmentation or topic modeling, with no prior labeling.
Reinforcement learning excels in environments where an agent must make sequential decisions and adapt to
feedback, such as robotics, gaming, or self-driving cars.
Each approach has distinct strengths: supervised learning for accurate predictions,
unsupervised learning for discovery, and reinforcement learning for complex adaptive behaviors.
