MPSeDC GET Program 3.0 2026 – Exam-Oriented Emerging Technologies(AI | Machine Learning | Blockchain | Quantum Computing) MCQ Questions
BEGINNER LEVEL
Q1: What does AI stand for?
A) Automatic Information
B) Artificial Intelligence
C) Advanced Integration
D) Auto Interaction
Correct Answer: B
Explanation: AI = Artificial Intelligence, the simulation of human intelligence by machines.
Exam Trick: Remember: AI = ‘Artificial’ (man-made) + ‘Intelligence’ (thinking).
Q2: Who is known as the ‘Father of Artificial Intelligence’?
A) Alan Turing
B) John McCarthy
C) Geoffrey Hinton
D) Marvin Minsky
Correct Answer: B
Explanation: John McCarthy coined the term ‘Artificial Intelligence’ in 1956 at the Dartmouth Conference.
Exam Trick: McCarthy = Coined term AI in 1956 Dartmouth.
Q3: Machine Learning is a subset of:
A) Blockchain
B) Artificial Intelligence
C) Quantum Computing
D) Cloud Computing
Correct Answer: B
Explanation: ML is a branch of AI that enables systems to learn from data without explicit programming.
Exam Trick: AI is the big circle; ML is inside it; DL is inside ML.
Q4: Which of the following is NOT a type of Machine Learning?
A) Supervised Learning
B) Unsupervised Learning
C) Reinforcement Learning
D) Sequential Learning
Correct Answer: D
Explanation: The three main ML types are Supervised, Unsupervised, and Reinforcement Learning.
Exam Trick: SUR = Supervised, Unsupervised, Reinforcement.
Q5: In Supervised Learning, the dataset used is:
A) Unlabeled
B) Labeled
C) Random
D) Encrypted
Correct Answer: B
Explanation: Supervised learning uses labeled data (input-output pairs) to train models.
Exam Trick: Supervised = Teacher gives labeled answers.
Q6: In Unsupervised Learning, the dataset is:
A) Labeled
B) Labeled and Unlabeled
C) Unlabeled
D) Encrypted
Correct Answer: C
Explanation: Unsupervised learning finds patterns in unlabeled data (e.g., clustering).
Exam Trick: Un-supervised = No teacher = No labels.
Q7: Which algorithm is commonly used for clustering?
A) Linear Regression
B) K-Means
C) Decision Tree
D) Logistic Regression
Correct Answer: B
Explanation: K-Means is a popular unsupervised clustering algorithm.
Exam Trick: K-Means = K Means ‘Kitne clusters’ (how many clusters).
Q8: Deep Learning is based on the structure of:
A) Computer circuits
B) Human neurons (Neural Networks)
C) Database tables
D) Blockchain blocks
Correct Answer: B
Explanation: Deep Learning uses Artificial Neural Networks inspired by the human brain.
Exam Trick: Deep Learning = Neural Network = Brain-like layers.
Q9: Blockchain is best described as a:
A) Centralized database
B) Distributed/Decentralized ledger
C) Single server storage
D) Cloud backup tool
Correct Answer: B
Explanation: Blockchain is a decentralized, distributed ledger maintained across multiple nodes.
Exam Trick: Blockchain = Chain of Blocks = No single owner.
Q10: Bitcoin is based on which technology?
A) Cloud Computing
B) Blockchain
C) Quantum Computing
D) Artificial Intelligence
Correct Answer: B
Explanation: Bitcoin, the first cryptocurrency, runs on blockchain technology introduced in 2008/2009.
Exam Trick: Bitcoin = Blockchain’s first big use case.
Q11: A ‘Block’ in blockchain primarily contains:
A) Only images
B) Transaction data, hash, previous hash
C) Only video files
D) Only user passwords
Correct Answer: B
Explanation: Each block contains transaction data, its own hash, and the previous block’s hash.
Exam Trick: Block = Data + Hash + Previous Hash (chain link).
Q12: Who introduced Bitcoin and Blockchain concept?
A) Vitalik Buterin
B) Satoshi Nakamoto
C) Tim Berners-Lee
D) Elon Musk
Correct Answer: B
Explanation: Satoshi Nakamoto published the Bitcoin whitepaper in 2008 (pseudonymous identity).
Exam Trick: Satoshi = Creator of Bitcoin/Blockchain whitepaper 2008.
Q13: Blockchain data, once added, is:
A) Easily editable
B) Immutable (cannot be easily altered)
C) Automatically deleted
D) Visible only to admin
Correct Answer: B
Explanation: A key property of blockchain is immutability — data cannot be changed once confirmed.
Exam Trick: Immutable = ‘I’m-mutable’ NO, cannot mutate!
Q14: The basic unit of information in Quantum Computing is called:
A) Bit
B) Byte
C) Qubit
D) Pixel
Correct Answer: C
Explanation: Quantum computers use qubits (quantum bits) instead of classical bits.
Exam Trick: Qubit = Quantum + Bit.
Q15: A classical bit can be in state:
A) 0 or 1 only
B) Both 0 and 1 simultaneously
C) Only 1
D) Only undefined
Correct Answer: A
Explanation: Classical bits are binary: strictly 0 OR 1.
Exam Trick: Classical = Either/Or (strict).
Q16: A qubit can exist in:
A) 0 or 1 only
B) Superposition of 0 and 1 simultaneously
C) Only negative values
D) Only decimal values
Correct Answer: B
Explanation: Qubits leverage superposition, allowing them to represent 0 and 1 simultaneously.
Exam Trick: Qubit = Superposition = Both at once.
Q17: Quantum Computing is based on principles of:
A) Classical Mechanics
B) Quantum Mechanics
C) Thermodynamics
D) Newtonian Physics
Correct Answer: B
Explanation: Quantum computing exploits quantum mechanical phenomena like superposition and entanglement.
Exam Trick: Quantum Computing = Quantum Mechanics (obvious link).
Q18: Which company built one of the first practical quantum computers?
A) Facebook
B) IBM
C) Twitter
D) Adobe
Correct Answer: B
Explanation: IBM has been a pioneer in building and providing access to quantum computers (IBM Q).
Exam Trick: IBM = Pioneer brand in Quantum Computing.
Q19: AI applications include all EXCEPT:
A) Chatbots
B) Self-driving cars
C) Manual paper filing
D) Voice assistants
Correct Answer: C
Explanation: Manual paper filing is a traditional non-AI process; the others use AI techniques.
Exam Trick: Find the ‘odd one out’ — manual = no AI.
Q20: Examples of AI-powered virtual assistants include:
A) Siri, Alexa, Google Assistant
B) MS Word, Excel, PowerPoint
C) Chrome, Firefox, Edge
D) Windows, Linux, macOS
Correct Answer: A
Explanation: Siri, Alexa, and Google Assistant use AI/NLP to understand and respond to users.
Exam Trick: SAG = Siri, Alexa, Google Assistant (AI assistants).
Q21: Natural Language Processing (NLP) deals with:
A) Image recognition
B) Human language understanding by machines
C) Hardware design
D) Network routing
Correct Answer: B
Explanation: NLP enables machines to understand, interpret, and generate human language.
Exam Trick: NLP = Language (Natural Language) Processing.
Q22: Computer Vision is a field of AI related to:
A) Understanding images/videos
B) Understanding sound only
C) Understanding text only
D) Understanding currency
Correct Answer: A
Explanation: Computer Vision enables machines to interpret and process visual data (images/videos).
Exam Trick: Vision = Eyes = Images/Video.
Q23: Which of these is a real-world Blockchain application besides cryptocurrency?
A) Supply chain management
B) Word processing
C) Video editing
D) Audio streaming
Correct Answer: A
Explanation: Blockchain is used in supply chain for transparent, tamper-proof tracking of goods.
Exam Trick: Blockchain uses: Crypto, Supply Chain, Voting, Smart Contracts.
Q24: A ‘Smart Contract’ is:
A) A paper-based legal document
B) Self-executing code on blockchain
C) An AI chatbot
D) A quantum algorithm
Correct Answer: B
Explanation: Smart contracts are self-executing programs stored on a blockchain that run when conditions are met.
Exam Trick: Smart Contract = Code that auto-executes (no middleman).
Q25: Ethereum is mainly known for introducing:
A) Quantum bits
B) Smart Contracts on blockchain
C) Neural Networks
D) 5G Networks
Correct Answer: B
Explanation: Ethereum, launched by Vitalik Buterin, popularized smart contracts on blockchain.
Exam Trick: Ethereum = Smart Contract Platform.
Q26: Training data in Machine Learning is used to:
A) Test final accuracy only
B) Teach the model patterns
C) Delete the model
D) Encrypt the dataset
Correct Answer: B
Explanation: Training data is fed into the algorithm so the model can learn underlying patterns.
Exam Trick: Training data = ‘Teacher’s notes’ for the model.
Q27: Which is an example of a Reinforcement Learning application?
A) Spam email filter
B) Game-playing AI (e.g., AlphaGo)
C) Spreadsheet formula
D) Static webpage
Correct Answer: B
Explanation: Reinforcement learning trains agents via rewards/penalties, used famously in AlphaGo.
Exam Trick: RL = Reward-based Learning = Games/Robots.
Q28: ‘Hash function’ in blockchain is used to:
A) Encrypt and uniquely identify block data
B) Delete blocks
C) Slow down transactions
D) Store images only
Correct Answer: A
Explanation: A hash function converts block data into a fixed-size unique digital fingerprint.
Exam Trick: Hash = Unique Digital Fingerprint.
Q29: Quantum Computing primarily promises faster solutions for:
A) Typing speed
B) Complex computational problems (e.g., cryptography, optimization)
C) Internet browsing speed
D) Printer speed
Correct Answer: B
Explanation: Quantum computers can theoretically solve certain complex problems exponentially faster than classical computers.
Exam Trick: Quantum = Speed for HARD problems, not daily tasks.
INTERMEDIATE LEVEL
Q30: Which term refers to a computer’s ability to improve performance through experience without explicit programming?
A) Hardcoding
B) Machine Learning
C) Manual Scripting
D) Static Programming
Correct Answer: B
Explanation: This is the classic definition of Machine Learning (Arthur Samuel, 1959).
Exam Trick: Improves with experience = Learning, not coding.
Q31: Which type of ML algorithm is used when output labels are continuous numeric values?
A) Classification
B) Regression
C) Clustering
D) Association
Correct Answer: B
Explanation: Regression predicts continuous numeric outputs (e.g., price prediction).
Exam Trick: Regression = Real numbers (continuous).
Q32: Which type of ML algorithm is used when output is a discrete category?
A) Regression
B) Classification
C) Dimensionality Reduction
D) Association Rule
Correct Answer: B
Explanation: Classification predicts discrete class labels (e.g., spam/not spam).
Exam Trick: Classification = Classes/Categories.
Q33: Which of these is a supervised learning algorithm?
A) K-Means Clustering
B) Decision Tree
C) Apriori Algorithm
D) PCA
Correct Answer: B
Explanation: Decision Tree is a supervised algorithm used for classification/regression with labeled data.
Exam Trick: K-Means & PCA & Apriori = Unsupervised; Decision Tree = Supervised.
Q34: Overfitting in Machine Learning means:
A) Model performs well on training but poorly on new/test data
B) Model performs poorly on training data only
C) Model is too simple
D) Model has no training data
Correct Answer: A
Explanation: Overfitting occurs when a model learns noise/details too specifically, hurting generalization.
Exam Trick: Overfit = ‘Over’-memorize, fails on new data.
Q35: Underfitting in Machine Learning means:
A) Model is too complex
B) Model is too simple to capture patterns
C) Model has 100% accuracy
D) Model uses too much data
Correct Answer: B
Explanation: Underfitting happens when a model is too simple to learn the underlying pattern of data.
Exam Trick: Underfit = ‘Under’-learned, too simple.
Q36: A bank wants to predict whether a loan applicant will default (Yes/No). Which ML approach fits best?
A) Regression
B) Classification
C) Clustering
D) Reinforcement Learning
Correct Answer: B
Explanation: Default Yes/No is a binary discrete outcome — a classification problem.
Exam Trick: Yes/No output = Classification.
Q37: An e-commerce company wants to group customers with similar buying behavior without predefined categories. Best approach?
A) Classification
B) Clustering (unsupervised)
C) Regression
D) Smart Contract
Correct Answer: B
Explanation: Grouping without predefined labels is a classic unsupervised clustering task.
Exam Trick: No labels + Grouping = Clustering.
Q38: Which of these is NOT a popular ML library/framework?
A) TensorFlow
B) PyTorch
C) Scikit-learn
D) Solidity
Correct Answer: D
Explanation: Solidity is a programming language for Ethereum smart contracts, not an ML library.
Exam Trick: Solidity = Blockchain (Ethereum), not ML.
Q39: Which language is primarily used to write Ethereum smart contracts?
A) Python
B) Solidity
C) Java
D) COBOL
Correct Answer: B
Explanation: Solidity is the dominant smart contract language for the Ethereum blockchain.
Exam Trick: Solidity = Ethereum’s smart contract language.
Q40: Consensus mechanism used by Bitcoin is:
A) Proof of Stake
B) Proof of Work
C) Proof of Authority
D) Proof of Capacity
Correct Answer: B
Explanation: Bitcoin uses Proof of Work (PoW), requiring miners to solve computational puzzles.
Exam Trick: Bitcoin = PoW (energy-intensive mining).
Q41: Ethereum shifted from Proof of Work to which consensus mechanism (post ‘The Merge’)?
A) Proof of Stake
B) Proof of Burn
C) Proof of Elapsed Time
D) Proof of Space
Correct Answer: A
Explanation: Ethereum moved to Proof of Stake (PoS) in 2022 (‘The Merge’) to reduce energy use.
Exam Trick: Ethereum Merge 2022 = PoW to PoS.
Q42: In blockchain, 51% attack refers to:
A) A discount on transaction fees
B) An entity controlling majority of network’s mining/validation power
C) A software update
D) A type of encryption
Correct Answer: B
Explanation: A 51% attack happens when a single entity controls majority hashing power, enabling double-spending.
Exam Trick: 51% = Majority Control = Attack risk.
Q43: Public blockchains (like Bitcoin) are characterized by:
A) Permissioned, restricted access
B) Permissionless, open to anyone
C) Single organization control
D) No transaction history
Correct Answer: B
Explanation: Public blockchains are open/permissionless — anyone can join, view, and validate.
Exam Trick: Public = Open to all = Permissionless.
Q44: Private/Permissioned blockchains are typically used by:
A) Anonymous public users only
B) Enterprises/consortiums needing controlled access
C) No one
D) Only individual hobbyists
Correct Answer: B
Explanation: Private blockchains restrict access to authorized participants, common in enterprise use (e.g., Hyperledger).
Exam Trick: Private Blockchain = Enterprise + Controlled access.
Q45: Hyperledger Fabric is an example of:
A) Public Blockchain
B) Permissioned/Private Blockchain framework
C) Quantum Algorithm
D) ML Framework
Correct Answer: B
Explanation: Hyperledger Fabric, by Linux Foundation, is a permissioned blockchain framework for enterprises.
Exam Trick: Hyperledger = IBM/Linux Foundation enterprise blockchain.
Q46: Which quantum principle allows two qubits to be correlated regardless of distance?
A) Superposition
B) Entanglement
C) Decoherence
D) Tunneling
Correct Answer: B
Explanation: Quantum entanglement links qubits such that the state of one instantly relates to the other, regardless of distance.
Exam Trick: Entanglement = ‘Entangled twins’ — linked at a distance.
Q47: ‘Quantum Supremacy’ refers to:
A) Quantum computers replacing classical ones entirely
B) A quantum computer solving a problem infeasible for classical computers
C) Quantum computers being cheaper than classical
D) A type of encryption
Correct Answer: B
Explanation: Quantum supremacy is achieving a computation a classical computer cannot feasibly perform.
Exam Trick: Supremacy = Quantum WINS where classical FAILS.
Q48: Google claimed quantum supremacy in which year with its Sycamore processor?
A) 2015
B) 2019
C) 2021
D) 2023
Correct Answer: B
Explanation: Google announced quantum supremacy in 2019 using its 53-qubit Sycamore processor.
Exam Trick: Google Sycamore = 2019 Quantum Supremacy claim.
Q49: Which algorithm shows quantum computers could break classical RSA encryption efficiently?
A) Grover’s Algorithm
B) Shor’s Algorithm
C) Dijkstra’s Algorithm
D) Apriori Algorithm
Correct Answer: B
Explanation: Shor’s Algorithm can factor large numbers exponentially faster, threatening RSA encryption.
Exam Trick: Shor’s = Security threat (factoring/RSA).
Q50: Grover’s Algorithm provides a quantum speedup for:
A) Sorting
B) Unstructured search problems
C) Image compression
D) Blockchain mining only
Correct Answer: B
Explanation: Grover’s algorithm offers quadratic speedup for searching unsorted databases.
Exam Trick: Grover = Search faster (quadratic speedup).
Q51: A model is trained on 1000 labeled images of cats/dogs and tested on new images to predict the class. This is:
A) Unsupervised Learning
B) Supervised Classification
C) Reinforcement Learning
D) Quantum Learning
Correct Answer: B
Explanation: Labeled training data + discrete class prediction = Supervised Classification.
Exam Trick: Labeled + Class prediction = Supervised Classification.
Q52: Identify the error: ‘K-Means is a Supervised Learning algorithm used for classification.’
A) No error
B) K-Means is Unsupervised, used for clustering, not classification
C) K-Means works only on text
D) K-Means requires labeled data
Correct Answer: B
Explanation: K-Means is an unsupervised clustering algorithm; it does not use labeled data for classification.
Exam Trick: K-Means = Unsupervised + Clustering, NOT supervised/classification.
Q53: Identify the error: ‘Blockchain data can be easily edited by any single node without consensus.’
A) No error
B) Blockchain requires network consensus; single node cannot alter validated data
C) Blockchain has no nodes
D) Blockchain is centralized
Correct Answer: B
Explanation: Blockchain’s strength is immutability via distributed consensus — no single node can unilaterally alter data.
Exam Trick: Blockchain = Needs CONSENSUS to change, not single-node edits.
Q54: Which of these is an example of Generative AI?
A) Linear Regression model predicting house price
B) ChatGPT generating text responses
C) K-Means clustering customers
D) Decision Tree classifying loans
Correct Answer: B
Explanation: Generative AI creates new content (text, images); ChatGPT generates human-like text.
Exam Trick: Generative AI = CREATES new content (text/image).
Q55: GPT stands for:
A) General Processing Tool
B) Generative Pre-trained Transformer
C) Global Processing Technique
D) Graphical Programming Tool
Correct Answer: B
Explanation: GPT = Generative Pre-trained Transformer, the architecture behind models like ChatGPT.
Exam Trick: GPT = Generative + Pre-trained + Transformer.
Q56: Which neural network architecture is widely used for image recognition?
A) RNN
B) CNN (Convolutional Neural Network)
C) Blockchain Network
D) Quantum Circuit
Correct Answer: B
Explanation: CNNs are specialized for processing grid-like data such as images, using convolution layers.
Exam Trick: CNN = Convolutional = Computer Vision/Images.
Q57: Which neural network architecture is best suited for sequential data like text/time-series?
A) CNN
B) RNN (Recurrent Neural Network)
C) K-Means
D) Apriori
Correct Answer: B
Explanation: RNNs have memory of previous inputs, making them suitable for sequences like text/time-series.
Exam Trick: RNN = Recurrent = Remembers sequence/time.
Q58: A logistics company wants to track goods transparently across multiple suppliers, preventing fraud. Best technology?
A) Quantum Computing
B) Blockchain
C) Regression Model
D) Reinforcement Learning
Correct Answer: B
Explanation: Blockchain provides a transparent, tamper-resistant, shared ledger ideal for supply chain tracking.
Exam Trick: Multi-party transparency/fraud prevention = Blockchain.
Q59: A pharma company needs to simulate molecular interactions for new drug discovery faster than classical computers allow. Best fit?
A) Blockchain
B) Quantum Computing
C) Basic spreadsheet
D) Linear Regression
Correct Answer: B
Explanation: Quantum computing excels at simulating quantum-level molecular interactions, valuable in drug discovery.
Exam Trick: Molecular simulation = Quantum Computing’s strength.
Q60: Which of these best defines ‘Decentralization’ in blockchain?
A) Control by a single central authority
B) Distribution of control across multiple nodes/participants
C) No control at all
D) Government-only control
Correct Answer: B
Explanation: Decentralization means no single entity controls the network; control is distributed among nodes.
Exam Trick: Decentralized = Power spread across many, not one.
Q61: A ‘Wallet’ in cryptocurrency context stores:
A) Physical cash
B) Public/Private keys to access digital assets
C) Only transaction history
D) Internet cookies
Correct Answer: B
Explanation: A crypto wallet stores the public/private key pair used to access and manage digital assets.
Exam Trick: Wallet = Keys (Public + Private), not actual coins physically.
Q62: Which term describes AI systems designed to perform a single, narrow task (e.g., facial recognition)?
A) General AI
B) Narrow AI (Weak AI)
C) Super AI
D) Quantum AI
Correct Answer: B
Explanation: Narrow AI (Weak AI) is designed for specific tasks; most current AI falls in this category.
Exam Trick: Narrow AI = ONE task only (most of today’s AI).
Q63: AI that can perform any intellectual task a human can is called:
A) Narrow AI
B) Artificial General Intelligence (AGI)
C) Reinforcement AI
D) Symbolic AI
Correct Answer: B
Explanation: AGI refers to hypothetical AI with human-level cognitive abilities across all domains.
Exam Trick: AGI = General = Human-level across ALL tasks (still theoretical).
Q64: Why is data preprocessing important in Machine Learning?
A) It is not important
B) It improves data quality and model accuracy
C) It deletes the dataset
D) It only formats colors
Correct Answer: B
Explanation: Preprocessing (cleaning, normalizing data) removes noise/inconsistencies, improving model performance.
Exam Trick: Garbage in = Garbage out; preprocessing avoids this.
Q65: A model gives 99% accuracy on training data but only 60% on test data. What is the likely issue?
A) Underfitting
B) Overfitting
C) Perfect model
D) Data leakage prevention
Correct Answer: B
Explanation: Large gap between training and test accuracy indicates the model has overfit the training data.
Exam Trick: Big gap train vs test = Overfitting signal.
Q66: Which of the following is a Quantum Computing hardware approach?
A) Superconducting qubits
B) HTML rendering engine
C) Relational database engine
D) TCP/IP stack
Correct Answer: A
Explanation: Superconducting qubits (used by IBM, Google) are a leading physical implementation of quantum computers.
Exam Trick: Superconducting qubits = IBM/Google hardware approach.
Q67: NFT (Non-Fungible Token) is built on which technology?
A) Quantum Computing
B) Blockchain
C) Reinforcement Learning
D) 5G
Correct Answer: B
Explanation: NFTs are unique digital assets recorded and verified on a blockchain.
Exam Trick: NFT = Unique digital asset on Blockchain.
Q68: Which statement about quantum bits (qubits) is FALSE?
A) Qubits can be in superposition
B) Qubits always behave exactly like classical bits
C) Qubits can be entangled
D) Measuring a qubit collapses its state
Correct Answer: B
Explanation: Qubits differ fundamentally from classical bits due to superposition and entanglement.
Exam Trick: Find FALSE: Qubits are NOT the same as classical bits in behavior.
Q69: Which of these best represents the relationship: AI contains ML contains DL?
A) Deep Learning contains Machine Learning which contains AI
B) AI contains Machine Learning which contains Deep Learning
C) All three are unrelated fields
D) ML contains AI which contains DL
Correct Answer: B
Explanation: AI is the broadest field, ML is a subset of AI, and Deep Learning is a subset of ML.
Exam Trick: AI > ML > DL (nested circles, biggest to smallest).
ADVANCED LEVEL
Q70: In Reinforcement Learning, the ‘Agent’ learns by interacting with the ‘Environment’ to maximize:
A) Loss function
B) Cumulative Reward
C) Number of epochs
D) Block size
Correct Answer: B
Explanation: RL agents take actions in an environment to maximize long-term cumulative reward through trial and error.
Exam Trick: RL = Agent + Environment + maximize Reward.
Q71: Which of the following best explains ‘Gradient Descent’ in Machine Learning?
A) A clustering technique
B) An optimization algorithm to minimize the loss function
C) A blockchain consensus method
D) A quantum algorithm
Correct Answer: B
Explanation: Gradient Descent iteratively adjusts model parameters to minimize the error/loss function.
Exam Trick: Gradient Descent = Walk DOWN the slope to minimize loss.
Q72: In a Neural Network, the ‘Activation Function’ is used to:
A) Store data permanently
B) Introduce non-linearity into the model
C) Encrypt the weights
D) Validate blockchain blocks
Correct Answer: B
Explanation: Activation functions (ReLU, Sigmoid, etc.) introduce non-linearity, enabling networks to learn complex patterns.
Exam Trick: Activation = Adds non-linearity = Learns complex patterns.
Q73: Which activation function outputs values between 0 and 1, commonly used for binary classification output layers?
A) ReLU
B) Sigmoid
C) Tanh
D) Softmax (multi-class)
Correct Answer: B
Explanation: Sigmoid squashes output between 0 and 1, ideal for binary classification probability output.
Exam Trick: Sigmoid = 0 to 1 = Probability for binary class.
Q74: ReLU (Rectified Linear Unit) activation function is defined as:
A) max(0, x)
B) 1/(1+e^-x)
C) tanh(x)
D) x^2
Correct Answer: A
Explanation: ReLU outputs the input directly if positive, otherwise zero: f(x) = max(0, x).
Exam Trick: ReLU = max(0,x): negative becomes 0, positive stays.
Q75: Given a Decision Tree splitting on ‘Income > 50000’, if applicant income = 45000, which branch is taken?
A) True branch (Income > 50000)
B) False branch (Income <= 50000)
C) Both branches
D) Neither, error occurs
Correct Answer: B
Explanation: Since 45000 is not greater than 50000, the condition is False, so the False branch is taken.
Exam Trick: 45000 < 50000 = condition False = False branch.
Q76: A classification model has high precision but low recall. What does this mean?
A) Model rarely predicts positive, but when it does, it’s mostly correct; misses many actual positives
B) Model predicts everything as positive
C) Model has zero errors
D) Model recall and precision are same thing
Correct Answer: A
Explanation: High precision/low recall means few false positives but many false negatives (missed positives).
Exam Trick: High Precision = Accurate positives; Low Recall = Misses positives.
Q77: F1-Score is the harmonic mean of:
A) Accuracy and Loss
B) Precision and Recall
C) Bias and Variance
D) Mean and Median
Correct Answer: B
Explanation: F1-Score balances Precision and Recall using their harmonic mean, useful for imbalanced datasets.
Exam Trick: F1 = Harmonic mean of Precision & Recall.
Q78: ‘Bias-Variance Tradeoff’ in ML refers to:
A) Balancing model simplicity (bias) vs sensitivity to data (variance)
B) Balancing CPU vs GPU usage
C) Balancing public vs private blockchain
D) Balancing qubits vs classical bits
Correct Answer: A
Explanation: This tradeoff balances underfitting (high bias) against overfitting (high variance) to optimize generalization.
Exam Trick: Bias = too simple; Variance = too sensitive; balance both.
Q79: In blockchain, a ‘Merkle Tree’ is used to:
A) Store images
B) Efficiently and securely verify large sets of transaction data
C) Mine new coins
D) Encrypt private keys only
Correct Answer: B
Explanation: A Merkle Tree hashes transaction pairs hierarchically, enabling efficient and secure data verification.
Exam Trick: Merkle Tree = Tree of Hashes = Efficient verification.
Q80: A blockchain network experiences a fork where two valid chains temporarily exist. This is called:
A) Merkle conflict
B) Blockchain Fork
C) Quantum Collapse
D) Gradient Explosion
Correct Answer: B
Explanation: A fork occurs when the blockchain splits into two potential paths, often resolved by the longest-chain rule.
Exam Trick: Fork = Chain splits into two paths temporarily.
Q81: ‘Double-spending problem’ in cryptocurrency is solved primarily by:
A) Centralized banks
B) Blockchain’s distributed consensus mechanism
C) Email verification
D) Manual auditing
Correct Answer: B
Explanation: Blockchain consensus ensures each coin can only be spent once, validated by the network.
Exam Trick: Double-spend = Solved by Consensus (network agreement).
Q82: Which quantum computing concept states that observing/measuring a qubit forces it into a definite state (0 or 1)?
A) Superposition
B) Entanglement
C) Wavefunction Collapse (Decoherence)
D) Quantum Tunneling
Correct Answer: C
Explanation: Measurement causes the qubit’s superposition to collapse into one definite classical state.
Exam Trick: Measure qubit = Superposition COLLAPSES to 0 or 1.
Q83: No-Cloning Theorem in Quantum Computing states:
A) Qubits can be copied perfectly
B) An arbitrary unknown quantum state cannot be copied exactly
C) Classical bits cannot be copied
D) Blockchain blocks cannot be duplicated
Correct Answer: B
Explanation: The no-cloning theorem proves that an unknown quantum state cannot be perfectly duplicated, vital for quantum cryptography.
Exam Trick: No-Cloning = Can’t perfectly copy unknown quantum state.
Q84: Quantum Key Distribution (QKD) primarily provides:
A) Faster internet speed
B) Theoretically unbreakable secure communication using quantum mechanics
C) Cheaper data storage
D) Blockchain mining acceleration
Correct Answer: B
Explanation: QKD uses quantum mechanics (e.g., BB84 protocol) to detect eavesdropping, enabling provably secure key exchange.
Exam Trick: QKD = Quantum-secured key exchange (detects eavesdropping).
Q85: Explain why Quantum Computers are not expected to fully ‘replace’ classical computers.
A) Quantum computers are cheaper for all tasks
B) Quantum computers excel only at specific problem types; classical computers remain efficient for general tasks
C) Quantum computers cannot perform any computation
D) Classical computers will be banned
Correct Answer: B
Explanation: Quantum computers offer advantage for specific problems (optimization, factoring, simulation), not general-purpose computing.
Exam Trick: Quantum = Specialist tool, not a general replacement.
Q86: Identify the error: ‘Proof of Stake requires miners to solve complex puzzles using high computational power like Bitcoin.’
A) No error
B) That describes Proof of Work, not Proof of Stake; PoS uses staked coins to validate
C) PoS uses no validators
D) PoS is identical to PoW
Correct Answer: B
Explanation: Proof of Stake selects validators based on staked cryptocurrency, not computational puzzle-solving (that’s PoW).
Exam Trick: PoW = computation/mining; PoS = staking coins (different!).
Q87: A blockchain transaction is rejected because the hash of the previous block doesn’t match. This indicates:
A) Successful transaction
B) Possible tampering/corruption of blockchain data
C) Normal mining process
D) Increased block reward
Correct Answer: B
Explanation: Mismatched previous-block hash signals data tampering, breaking the chain’s integrity.
Exam Trick: Hash mismatch = Tampering detected = Chain broken.
Q88: Which of these is a real limitation of current (NISQ-era) Quantum Computers?
A) Too cheap to build
B) High error rates and quantum decoherence
C) Too many qubits already perfected
D) No use cases at all
Correct Answer: B
Explanation: Current quantum computers (Noisy Intermediate-Scale Quantum) suffer from noise/errors and short coherence times.
Exam Trick: NISQ = Noisy = Error-prone, still maturing technology.
Q89: Which ML technique reduces the number of input features while preserving important information?
A) Clustering
B) Dimensionality Reduction (e.g., PCA)
C) Regression
D) Classification
Correct Answer: B
Explanation: Dimensionality reduction (e.g., PCA) compresses features while retaining maximum variance/information.
Exam Trick: PCA = Principal Component Analysis = Reduce dimensions.
Q90: If a CNN’s convolution layer uses a 3×3 filter on a 5×5 image with stride 1 and no padding, what is the output size?
A) 5×5
B) 3×3
C) 2×2
D) 6×6
Correct Answer: B
Explanation: Output size = (Input – Filter)/Stride + 1 = (5-3)/1 + 1 = 3, so output is 3×3.
Exam Trick: Formula: (N-F)/S + 1 = (5-3)/1+1 = 3.
Q91: ‘Transfer Learning’ in Deep Learning refers to:
A) Training a model from scratch every time
B) Reusing a pre-trained model on a new related task
C) Transferring data between blockchains
D) Transferring qubits between computers
Correct Answer: B
Explanation: Transfer learning leverages a model already trained on one task to speed up learning on a related task.
Exam Trick: Transfer Learning = Reuse pre-trained knowledge.
Q92: Which of the following best describes a ‘51% attack’ impact on PoW blockchains?
A) Improves transaction speed only
B) Allows attacker to double-spend and reverse transactions
C) Has no real security impact
D) Increases decentralization
Correct Answer: B
Explanation: With majority hash power, an attacker can rewrite transaction history and double-spend coins.
Exam Trick: 51% attack = Majority power = Can rewrite history.
Q93: A hospital wants patient records to be tamper-proof, auditable, and shared securely among authorized doctors only. Best solution?
A) Public Blockchain open to all
B) Permissioned/Private Blockchain
C) Plain Excel sheet
D) Public cloud storage with no encryption
Correct Answer: B
Explanation: Permissioned blockchain restricts access to authorized parties while ensuring immutability and auditability.
Exam Trick: Sensitive + Controlled access + Tamper-proof = Permissioned Blockchain.
Q94: A financial firm needs to detect fraudulent transactions in real time from millions of records with complex patterns. Best approach?
A) Manual review
B) Machine Learning-based anomaly detection
C) Quantum Key Distribution
D) Simple if-else rules only
Correct Answer: B
Explanation: ML models (e.g., anomaly detection algorithms) can identify subtle fraud patterns at scale, better than manual or static rules.
Exam Trick: Pattern detection at scale = ML anomaly detection.
Q95: Which term refers to AI decision-making processes that are not easily interpretable by humans?
A) Explainable AI (XAI)
B) Black Box AI
C) Open Source AI
D) Transparent AI
Correct Answer: B
Explanation: ‘Black Box AI’ describes models (e.g., deep neural nets) whose internal decision logic is hard to interpret.
Exam Trick: Black Box = Can’t see inside = Hard to explain decisions.
Q96: Explainable AI (XAI) aims to:
A) Hide model decisions
B) Make AI decision-making transparent and interpretable
C) Increase model complexity only
D) Replace all classical computers
Correct Answer: B
Explanation: XAI focuses on making AI’s reasoning understandable to humans, important for trust and accountability.
Exam Trick: XAI = eXplainable = Transparent reasoning.
Q97: Why is ‘Quantum Entanglement’ considered useful but also fragile for computing?
A) It provides correlation across qubits but is highly sensitive to environmental noise (decoherence)
B) It has no use in computing
C) It makes qubits permanently stable
D) It only works in classical computers
Correct Answer: A
Explanation: Entanglement enables powerful correlations for computation/communication but is easily disrupted by external noise.
Exam Trick: Entanglement = Powerful but FRAGILE (noise breaks it).
Q98: Which statement is TRUE regarding Blockchain and Quantum Computing’s future relationship?
A) Quantum computing has no impact on blockchain security
B) Sufficiently powerful quantum computers could threaten current cryptographic algorithms used in blockchain
C) Blockchain makes quantum computers faster
D) Quantum computing will replace blockchain entirely
Correct Answer: B
Explanation: Quantum computers running Shor’s algorithm could break ECC/RSA-based blockchain cryptography, prompting ‘post-quantum cryptography’ research.
Exam Trick: Quantum threat to Blockchain = Breaking current encryption (Shor’s algo).
Q99: ‘Post-Quantum Cryptography’ refers to:
A) Cryptography used only in the past
B) New cryptographic algorithms designed to resist quantum computer attacks
C) Cryptography that requires no encryption
D) A type of blockchain consensus
Correct Answer: B
Explanation: Post-quantum cryptography develops algorithms resistant to attacks by quantum computers (e.g., lattice-based crypto).
Exam Trick: Post-Quantum Crypto = Quantum-resistant security algorithms.
Q100: An ML model’s loss value increases steadily during training instead of decreasing. Likely cause?
A) Perfect model convergence
B) Learning rate too high causing divergence
C) Too much training data
D) No issue at all
Correct Answer: B
Explanation: An excessively high learning rate can cause the optimizer to overshoot, making loss diverge instead of converge.
Exam Trick: Loss increasing = Learning rate likely too high.