
CBSE Class 10 Artificial Intelligence (AI) Quiz Questions & Exam
Welcome to the definitive resource for CBSE Class 10 Artificial Intelligence. Whether you are distinguishing between AI, Machine Learning, and Deep Learning, or analysing complex Confusion Matrices, this interactive platform bridges the gap between theoretical concepts and practical application.
Our curriculum-aligned modules dive deep into the three core domains—Data Science, Computer Vision, and Natural Language Processing (NLP)—while equipping you with essential Python skills (NumPy, Pandas, Matplotlib). From understanding the architecture of Convolutional Neural Networks (CNNs) to navigating the moral complexities of AI Ethics and Data Privacy, Rezyo ensures you are exam-ready. Test your knowledge with our adaptive quizzes and master the AI Project Cycle today.
Master
Intelligence.
A complete guide to Artificial Intelligence, Machine Learning, and Neural Networks. Updated for 2025.
What is Intelligence?
Intelligence is the ability to perceive information, retain it as knowledge, and apply it to solve problems within a specific environment. It involves traits like reasoning, planning, learning, and adaptation.
In machines, we call this Artificial Intelligence (AI). An AI machine mimics human traits: it makes decisions, predicts outcomes, and improves on its own based on data.
AI vs. Automation
A washing machine is automation, not AI. It follows a fixed set of instructions. An AI machine, like a self-driving car, makes decisions based on real-time data it hasn’t seen before.
The Hierarchy of AI
Figure 1: Relationship between AI, ML, and DL.
Comparison: AI, ML, and Deep Learning
| Term | Definition | Key Characteristic |
|---|---|---|
| Artificial Intelligence (AI) | Any technique enabling computers to mimic human intelligence. | Broad umbrella term. |
| Machine Learning (ML) | A subset of AI. Machines improve at tasks with experience (data). | Learn without explicit programming. |
| Deep Learning (DL) | A subset of ML. Uses neural networks to train itself on vast data. | Mimics the human brain (Neurons). |
Data Science
Works with numeric and alpha-numeric data. Systems collect data, maintain datasets, and derive meaning. Used in price comparison sites, banking fraud detection, and genomics.
Computer Vision
Enables machines to analyze visual information (images/videos). Processes include image acquiring, screening, and identifying. Used in self-driving cars, face locks, and medical imaging.
Natural Language Processing
Deals with interaction between computers and human language. Extracts meaning from spoken or written words. Used in email spam filters, chatbots, and voice assistants like Alexa.
Advanced Concepts
NLP: Text Processing Pipeline
Machines cannot read text directly. We must convert natural language into numbers using a pipeline.
- Sentence Segmentation: Dividing corpus into sentences.
- Tokenisation: Breaking sentences into individual words (tokens).
- Stopwords Removal: Removing common words (and, the, is) that add little meaning.
- Stemming: Cutting off affixes to get the root (e.g., “studies” -> “studi”). Fast but crude.
- Lemmatization: Reducing words to meaningful base forms (e.g., “studies” -> “study”). Slower but accurate.
- Bag of Words: Counts frequency of words. Ignores grammar/order. Creates a document vector.
- TFIDF: Term Frequency-Inverse Document Frequency. Assigns value to words. High value = Frequent in this doc but rare elsewhere (Important). Low value = Frequent everywhere (Common).
CV: Convolutional Neural Networks (CNN)
CNNs are deep learning algorithms designed to assign importance to various aspects of an image.
Extracts features like edges and corners. It uses a Kernel (a small matrix) that slides over the image to create a Feature Map.
Introduces non-linearity. It converts all negative pixel values to zero (black), making edges more distinct.
Reduces the size of the image while retaining important features. Max Pooling takes the largest number from a pixel group.
Flattened vector layer at the end that classifies the image into a label (e.g., “Car: 70%”).
DS: K-Nearest Neighbour (KNN)
A simple supervised learning algorithm for classification. It assumes similar things exist in close proximity.
How it works: To predict the label of a new point, we look at its ‘K’ nearest neighbours. If K=3, we check the 3 closest known points. If 2 are ‘Red’ and 1 is ‘Blue’, the new point is classified as ‘Red’.
AI Ethics
As AI becomes powerful, we must ensure it is safe, fair, and respects privacy.
Moral Issues: The Self-Driving Car
The Trolley Problem: If a self-driving car faces an unavoidable accident, should it hit a pedestrian to save the passenger, or swerve and hurt the passenger to save the pedestrian? The morals of the developer are transferred to the machine.
AI Bias
Machines learn from human data, which contains human biases.
Example: Most voice assistants (Siri, Alexa) default to female voices, reinforcing gender stereotypes.
Data Privacy
“Data is the new gold.”
- • Apps collect data via microphones, GPS, and cameras.
- • Data is often used for targeted advertising.
- • Consent is key: Users must know what they are sharing.
- • The Divide: AI access creates a gap between those who can afford smart devices and those who cannot, potentially leading to digital exclusion.
Python for AI
Python is the most popular language for AI due to its simplicity and vast library support.
NumPy
Fundamental package for numerical computations. Handles arrays and matrices faster than standard Python lists.
Pandas
Data manipulation library. Used for working with structured data (tables, CSVs) via DataFrames.
Matplotlib
Visualization library. Converts data into understandable graphs (Bar charts, Histograms, Scatter plots).
OpenCV
Open Source Computer Vision library. Processes images and videos (resizing, cropping, object detection).
Jupyter Notebook
An open-source web app that allows you to create and share documents containing live code, equations, visualizations, and narrative text. It is the standard environment for Data Science.
The AI Project Cycle
Building an AI project follows a structured framework. This ensures the solution is effective and relevant.
-
01.
Problem Scoping Setting the goal. Using the 4Ws Canvas (Who, What, Where, Why) to define the problem clearly.
-
02.
Data Acquisition Collecting authentic data from reliable sources (Surveys, Web Scraping, Sensors, API).
-
03.
Data Exploration Visualizing data (Graphs, Charts) to find patterns and trends.
-
04.
Modelling
Rule Based: Developer defines rules. Static.
Learning Based: Machine learns from data. Dynamic.
Supervised (Labelled data) vs Unsupervised (Unlabelled data). -
05.
Evaluation Testing the model on unseen data to calculate efficiency using metrics like Accuracy and F1 Score.
Model Evaluation
Evaluation helps us understand if our AI model is useful. We compare the model’s Prediction against the Reality.
The Confusion Matrix
Interactive F1 Score Analysis
Adjust values to see how Precision and Recall affect the F1 Score.
(TP + TN) / Total Cases
Percentage of correct predictions. Bad for imbalanced datasets (e.g., Forest Fires).
TP / (TP + FP)
Of all positives predicted, how many were actually positive? Important when False Positives are costly (e.g., Spam Filter).
TP / (TP + FN)
Of all actual positives, how many did we find? Important when False Negatives are dangerous (e.g., Viral Outbreak, Fire).
2 * (Precision * Recall) / (Precision + Recall)
Balance between Precision and Recall. High F1 Score indicates a good model.
Knowledge Check
Loading Question…
Quiz Complete!
You scored 0/50
Frequently Asked Questions
What is the 4Ws Problem Canvas?
It is a framework used in Problem Scoping. It consists of: Who (Stakeholders), What (Nature of problem), Where (Context/Location), and Why (Benefits of solution).
Is Deep Learning the same as AI?
No. Deep Learning is a specialized subset of Machine Learning, which itself is a subset of AI. Deep Learning uses neural networks to learn from vast amounts of data.
Why do we use F1 Score instead of Accuracy?
Accuracy can be misleading if the dataset is imbalanced (e.g., 98% of days have no fire). F1 Score balances Precision and Recall, giving a better measure of performance for critical cases.
What is a Neural Network?
A system modelled on the human brain. It consists of layers (Input, Hidden, Output) of nodes. It is able to automatically extract features without manual programming.




