CBSE Class 10 Artificial Intelligence, important questions & quick revision
A fast revision sheet for CBSE Class 10 Artificial Intelligence: 97+ important one-mark questions with their answers, grouped by topic. Use it for last-minute recall, then practise the full questions with step-by-step solutions.
Practise Artificial Intelligence free Take a mock test
Model Evaluation
- In a confusion matrix, a 'False Positive' means the model:Predicted positive, but the actual value was negative
- Precision is defined as:TP / (TP + FP)
- Recall (sensitivity) is defined as:TP / (TP + FN)
- Accuracy of a model is:(TP + TN) / (TP + TN + FP + FN)
- An AI model was tested on 1000 samples with TP=200, TN=600, FP=100, FN=100. How many predictions were correct?800
- For an AI model with TP=200, FP=100, FN=100, what is the Precision (to 2 decimals)?0.67
- An AI model was tested on 100 samples with TP=50, TN=30, FP=10, FN=10. How many predictions were correct?80
- For an AI model with TP=50, FP=10, FN=10, what is the Precision (to 2 decimals)?0.83
- An AI model was tested on 100 samples with TP=70, TN=20, FP=5, FN=5. How many predictions were correct?90
- For an AI model with TP=70, FP=5, FN=5, what is the Precision (to 2 decimals)?0.93
- An AI model was tested on 1000 samples with TP=150, TN=700, FP=100, FN=50. How many predictions were correct?850
- For an AI model with TP=150, FP=100, FN=50, what is the Precision (to 2 decimals)?0.6
Machine Learning
- Which scenario best represents a regression problem?Predicting tomorrow's temperature
- Classifying an email as 'spam' or 'not spam' is an example of:Classification
- Learning from data that has NO labels, by finding groups or patterns, is called:Unsupervised learning
- Grouping customers into segments without pre-defined labels is an example of:Clustering (unsupervised learning)
- Which learning approach is most suitable for training an AI model to park a car correctly through trial and error?Reinforcement learning
- Overfitting happens when a model:Memorises the training data instead of learning general patterns
- Why do we split a dataset into a training set and a test set?To train on one part and fairly check performance on unseen data
- In machine learning, 'error' is used to measure:How accurately the model predicts the data
- Learning from data that already has labelled answers is called:Supervised learning
- Grouping data without any labels is an example of:Unsupervised learning
- Learning by trial and error using rewards and penalties is called:Reinforcement learning
- Predicting a continuous number such as house price is an example of:Regression
Introduction to AI
- Which of the following best describes Artificial Intelligence?Machines performing tasks that usually need human intelligence
- The relationship between AI, Machine Learning and Deep Learning is best shown as:AI ⊃ ML ⊃ DL (DL is a subset of ML, which is a subset of AI)
- Which of these is one of the three domains of AI?Natural Language Processing
- An AI system that learns to play a game better by getting rewards and penalties is an example of:Reinforcement learning
- A machine that can only follow fixed rules written by a programmer and cannot learn is called:A rule-based AI
- Which AI domain would be MOST suitable for building a price-comparison website that reads product descriptions?Natural Language Processing
- Which AI domain deals with enabling machines to 'see' and interpret images?Computer Vision
- Which of the following is NOT an example of Artificial Intelligence?A basic calculator adding two numbers
- Deep Learning is a subset of:Machine Learning
- An AI system that recognises objects in photos belongs to the domain of:Computer Vision
- An AI that understands and replies to typed text belongs to the domain of:Natural Language Processing
- Which is a feature that makes a machine 'intelligent'?It can learn from data and improve
AI Project Cycle
- The first stage of the AI Project Cycle, where the goal of the project is defined, is called:Problem Scoping
- Collecting the data needed for an AI project is done in which stage?Data Acquisition
- The stage where the AI model is built and trained is called:Modelling
- Testing how well a trained model performs on new data is done in the ______ stage.Evaluation
- The '4Ws' problem canvas used in Problem Scoping stands for:Who, What, Where, Why
- Visualising data to find patterns before modelling is called:Data Exploration
- How many stages does the AI Project Cycle have?5
- The '4Ws' problem canvas is used in which stage of the AI Project Cycle?Problem Scoping
- Collecting the data needed for an AI project happens in the ______ stage.Data Acquisition
- Showing data as graphs and charts to understand it is part of:Data Exploration
- Checking how well the trained model performs is done in the ______ stage.Evaluation
Natural Language Processing
- NLP stands for:Natural Language Processing
- Converting all text to lowercase during text pre-processing helps because:'Hello' and 'hello' are then treated as the same word
- Reducing a word to its root by chopping off endings (which may give a non-dictionary word) is called:Stemming
- Reducing a word to its proper dictionary base form (its lemma) is called:Lemmatization
- The NLP model that represents text as a collection of word counts, ignoring grammar and order, is the:Bag of Words model
- To analyse customer reviews and judge whether they are positive or negative, the best NLP application is:Sentiment analysis
- Which type of chatbot requires coding and works directly on larger databases?Smart bot
- In the sentence 'Her face turned red with anger', the word 'red' shows which feature of natural language?Context-dependent meaning
- Breaking a sentence into individual words or tokens is called:Tokenisation
- Common words like 'the', 'is' and 'a' that are often removed in NLP are called:Stopwords
- Reducing words like 'playing' and 'played' to the root 'play' is called:Stemming
- A representation that counts how often each word appears in a text is the:Bag of Words
Employability Skills
- Which of the following is an imperative sentence?Shut the front door.
- Breaking down big goals into smaller parts makes the goal more ______.Achievable
- A strong password should consist of:Letters, numbers and special characters
- Which of the following is NOT a United Nations Sustainable Development Goal?Population
- The seven elements of effective communication are called the:7 C's of communication
- Self-management skill that lets you understand and manage your own emotions is called:Emotional intelligence
- ICT skills mainly help us to:Use technology for everyday tasks
- A key quality of a successful entrepreneur is:Willingness to take calculated risks
- Which of these is a good listening habit?Paying full attention to the speaker
- Setting a clear target you want to achieve by a certain date is called a:Goal
- Which key combination is commonly used to copy selected text?Ctrl + C
- An entrepreneur who creates a product that helps society and the environment shows:Social responsibility
AI Ethics
- Ethical frameworks in AI are mainly designed to:Ensure AI decisions do not cause unintended harm
- A framework built around the moral values of a particular society is a:Value-based ethical framework
- Bioethics, which deals with ethics in health, medicine and biology, is an example of a:Sector-based ethical framework
- When an AI system gives unfair results because its training data was not representative, this is called:AI bias
- Which of the following is the best practice for ethical AI?Using fair, representative and consented data
- When an AI system treats some groups unfairly because of its training data, this is called:AI bias
- Using AI to help diagnose diseases early is an example of:AI for good
- Taking a person's data without permission mainly violates their:Privacy
Neural Networks
- A neural network is made up of:Layers of interconnected nodes (neurons)
- In a neural network, the layer that receives the raw input data is the:Input layer
- The layers between the input and output layers of a neural network are called:Hidden layers
- CNN stands for:Convolutional Neural Network
- ANN stands for:Artificial Neural Network
- Which type of neural network is most commonly used for image recognition?Convolutional Neural Network (CNN)
- An artificial neural network is loosely inspired by the:Human brain
- The first layer that receives data in a neural network is the:Input layer
- Layers between the input and output layers of a neural network are called:Hidden layers
Computer Vision
- The smallest unit of a digital image is a:Pixel
- Which statement best describes the relationship between pixels and image resolution?The number of pixels in an image is known as its resolution
- A colour digital image is usually stored using which colour model?RGB (Red, Green, Blue)
- An AI task that identifies the type of an object in an image and assigns it a single label is called:Classification
- Drawing a bounding box to find WHERE an object is in an image is called:Localization
- In the context of Computer Vision vs Image Processing, which statement is correct?Computer Vision is a superset of Image Processing
- In the RGB colour model, the three channels are:Red, Green, Blue
- The number of pixels in an image (width times height) describes its:Resolution
- A grayscale pixel is usually stored as a value from:0 to 255
Turn revision into marks
Once you have skimmed these, attempt the full Artificial Intelligence topics with step-by-step solutions, then sit the Artificial Intelligence mock test to check pacing.
Related CBSE Class 10 subjects
Frequently asked questions
What are the most important questions for CBSE Class 10 Artificial Intelligence?
This page lists 97+ frequently-tested one-mark Artificial Intelligence questions with answers, grouped by topic, as a quick revision sheet.
How do I revise CBSE Class 10 Artificial Intelligence quickly?
Scan the one-line questions and answers by topic for recall, then attempt the full questions with solutions in the practice app to check understanding.
Are the Artificial Intelligence answers reliable?
Yes, every answer is checked against the marking scheme. If you spot an error, use the report control in the app.