Create a custom practice set
Pick category, difficulty, number of questions, and time limit. Start instantly with your own quiz.
Generate QuizGK Questions
Filter by category, type, and difficulty. Reading is open for everyone.
Which defense technique trains ML models with adversarial examples to improve robustness?
- Adversarial Training
- Defensive Distillation
- Randomized Smoothing
- All of these
Answer: All of these
Adversarial defenses: adversarial training includes perturbed examples, defensive distillation smooths outputs, randomized smoothing adds noise. All improve robustness against adversarial attacks.
Which Zero Trust principle requires continuous verification of trust assumptions?
- Never Trust, Always Verify
- Assume Breach
- Least Privilege Access
- All of these
Answer: All of these
Zero Trust combines: never trust always verify (continuous authentication), assume breach (detect/contain threats), least privilege (minimum access). All principles work together for comprehensive security.
Which architecture underpins modern LLMs?
- CNN
- RNN
- Transformer
- LSTM
Answer: Transformer
Transformer uses self-attention for parallel sequence processing.
Which method reduces LLM hallucinations by grounding responses in external knowledge?
- Fine-tuning
- RAG
- Prompt Engineering
- All of these
Answer: RAG
Retrieval-Augmented Generation combines LLMs with external knowledge retrieval.
The technique that enables efficient training of large models by partitioning across devices is called ________ Parallelism.
Answer: Model / Pipeline
Model parallelism splits model layers across devices; pipeline parallelism stages computation. Critical for scaling LLM training.
Which architecture forms the foundation of modern Large Language Models?
- CNN
- RNN
- Transformer
- LSTM
Answer: Transformer
Transformer architecture uses self-attention mechanisms to process sequences in parallel, enabling scaling to billions of parameters.
Which database design pattern stores historical changes to data by adding new rows instead of updating existing ones?
- Soft Delete
- Temporal Tables
- Event Sourcing
- Both B and C
Answer: Both B and C
Temporal Tables (SQL:2011 standard) automatically track row history with system-versioned tables. Event Sourcing stores state changes as immutable events, enabling reconstruction of any historical state. Both support audit trails, time-travel queries, and compliance. Trade-offs: increased storage and query complexity for historical analysis.
Zero Trust Architecture assumes that threats can exist both outside and inside the traditional network perimeter.
- True
- False
Answer: True
Zero Trust operates on 'never trust, always verify' principle. It eliminates implicit trust based on network location, requiring strict identity verification for every access request regardless of origin. Implements micro-segmentation, least-privilege access, continuous monitoring, and assumes breach mindset for robust security posture.
The cloud architecture pattern where applications are built as loosely coupled, independently deployable services is called ________ Architecture.
Answer: Microservices
Microservices Architecture decomposes applications into small, independent services communicating via APIs. Benefits: independent scaling, technology diversity, fault isolation, and faster deployments. Challenges: distributed system complexity, data consistency, monitoring. Often combined with containers (Docker) and orchestration (Kubernetes) for cloud-native development.
Which database isolation level prevents dirty reads but allows non-repeatable reads and phantom reads?
- Read Uncommitted
- Read Committed
- Repeatable Read
- Serializable
Answer: Read Committed
Read Committed isolation level ensures transactions see only committed data (prevents dirty reads). However, same query in same transaction may return different results if other transactions modify data (non-repeatable reads) or add new rows (phantom reads). Default in PostgreSQL and Oracle; balances consistency and concurrency.
Which ISRO mission demonstrated in-orbit servicing capability by refueling a satellite?
- SPADEX
- POEM
- DS-SAR
- Not yet demonstrated
Answer: Not yet demonstrated
As of 2026, ISRO has not yet demonstrated in-orbit refueling. SPADEX (Space Docking Experiment) tested docking/undocking, a prerequisite for servicing. POEM (PSLV Orbital Experimental Module) demonstrated using spent rocket stage as orbital platform. In-orbit servicing remains a future goal for satellite life extension and space debris management.
Homomorphic encryption allows computations on encrypted data without decrypting it first.
- True
- False
Answer: True
Homomorphic encryption enables performing calculations on ciphertext, producing encrypted results that, when decrypted, match operations performed on plaintext. Critical for privacy-preserving cloud computing, secure voting, and confidential AI inference. Still computationally intensive but advancing with research and hardware acceleration.
The framework that enables consent-based sharing of financial data between institutions in India is called the ________ Framework.
Answer: Account Aggregator
Account Aggregator (AA) is a consent-based data sharing framework regulated by RBI. It allows individuals to securely share financial information (bank statements, investments, insurance) between Financial Information Providers (FIPs) and Financial Information Users (FIUs) for better credit access and personalized services.
Which type of ransomware encrypts files and also exfiltrates data to threaten public release if ransom is not paid?
- Crypto-ransomware
- Locker ransomware
- Double extortion ransomware
- Scareware
Answer: Double extortion ransomware
Double extortion ransomware combines file encryption with data theft. Attackers threaten to publish sensitive data if ransom isn't paid, increasing pressure on victims. This evolved tactic makes backups insufficient as a sole defense. Prevention requires DLP, encryption, network segmentation, and incident response planning.
Which advantage do neuromorphic chips offer for edge AI?
- High Precision
- Ultra-Low Power
- Large Memory
- Fast Clock
Answer: Ultra-Low Power
Neuromorphic chips achieve ultra-low power via event-driven, sparse computation.
The technology that uses neural networks in biological tissue is called ________ Computing.
Answer: Wetware
Wetware computing uses biological neurons for computation. Early research stage.
Hash-based signatures are quantum-resistant but have larger signature sizes.
- True
- False
Answer: True
SPHINCS+ uses hash functions believed quantum-resistant. Trade-off: larger signatures.
The technique that enables training on encrypted data is called ________ Encryption.
Answer: Homomorphic
Homomorphic Encryption enables computation on ciphertext.
Which framework provides guidelines for securing IoT deployments?
- NIST IR 8259
- ETSI EN 303 645
- Both A and B
- Neither
Answer: Both A and B
NIST IR 8259 and ETSI EN 303 645 provide IoT security baselines: secure boot, updates, data protection.
Which DeFi risk arises from impermanent loss when providing liquidity to volatile trading pairs?
- Impermanent Loss
- Smart Contract Risk
- Oracle Manipulation
- All of these
Answer: Impermanent Loss
Impermanent loss occurs when token prices diverge, reducing value vs holding tokens. Critical risk for liquidity providers.