D684 Introduction to Computer Science - Set 4 - Part 3

Correct Answer: Live migration

Live migration transfers running virtual machines between servers while maintaining operation, enabling maintenance without service interruption.
Option a is incorrect because cold migration requires shutting down virtual machines, causing downtime during the transfer process.
Option c is incorrect because manual migration involves human intervention and downtime, disrupting services during maintenance periods.
Option d is incorrect because batch migration processes multiple machines together but doesn't maintain continuous operation during transfer.

Correct Answer: Speaker adaptation

Speaker adaptation fine-tunes recognition models to individual speech patterns and accents, improving accuracy for specific users.
Option a is incorrect because generic training produces average models that work poorly for users with unique accents or patterns.
Option c is incorrect because fixed models cannot improve with use, maintaining poor performance for non-standard speech patterns.
Option d is incorrect because offline training happens before deployment, unable to adapt to individual users' speaking characteristics.

Correct Answer: Behavior prediction

Behavior prediction models analyze pedestrian body language and movement patterns to anticipate actions, improving safety at crosswalks.
Option a is incorrect because rule-based systems cannot capture the complexity and variability of human pedestrian behavior patterns.
Option c is incorrect because random guessing provides no reliable safety improvement for pedestrian interaction in traffic.
Option d is incorrect because historical averaging doesn't account for individual pedestrian intentions or current contextual factors.

Correct Answer: Dynamic frequency scaling

Dynamic frequency scaling adjusts processor speed based on computational demand, optimizing battery life while maintaining responsive performance.
Option a is incorrect because maximum performance drains battery quickly regardless of actual computational needs or workload requirements.
Option c is incorrect because minimum power setting provides poor performance even when users need responsive computing power.
Option d is incorrect because fixed frequency cannot adapt to varying workloads, wasting power or providing insufficient performance.

Correct Answer: Input prediction

Input prediction anticipates likely player actions based on context, reducing perceived latency by starting responses before confirmation.
Option a is incorrect because input buffering accumulates commands, actually increasing latency rather than compensating for network delays.
Option c is incorrect because input queueing organizes commands but doesn't reduce the latency players perceive during gameplay.
Option d is incorrect because input delaying makes latency worse by adding artificial delays to input processing timing.

Correct Answer: Anomaly detection

Anomaly detection identifies deviations from normal patterns, catching unknown attack types that signature-based methods would miss.
Option a is incorrect because signature detection only recognizes known attack patterns, missing new or modified cyber threats.
Option c is incorrect because whitelist detection allows known good traffic but can't identify subtle attacks mimicking legitimate traffic.
Option d is incorrect because blacklist detection blocks known bad sources but misses attacks from new or compromised systems.

Correct Answer: Dead reckoning

Dead reckoning estimates position using last known location and vehicle speed/direction, maintaining navigation during signal loss.
Option a is incorrect because signal amplification cannot create GPS signals that don't exist inside tunnels or underground.
Option c is incorrect because random estimation would provide incorrect directions, potentially leading drivers dangerously off course.
Option d is incorrect because service suspension leaves drivers without guidance precisely when navigation is most needed.

Correct Answer: Machine learning

Machine learning analyzes user behavior patterns to identify personally important emails beyond simple rule-based spam filtering.
Option a is incorrect because static rules cannot adapt to individual preferences or evolving importance patterns over time.
Option c is incorrect because random filtering would mark important messages incorrectly, missing critical communications for users.
Option d is incorrect because time-based sorting assumes recent equals important, missing crucial older messages while highlighting unimportant new ones.

Correct Answer: Distributed control

Distributed control coordinates turbines as a system, optimizing collective power generation while respecting individual turbine safety limits.
Option a is incorrect because individual control ignores wake effects between turbines, reducing overall farm power generation efficiency.
Option c is incorrect because manual control cannot respond quickly enough to changing wind conditions across the farm.
Option d is incorrect because timed control ignores actual wind conditions, missing optimization opportunities and risking turbine damage.

Correct Answer: Cookie isolation

Cookie isolation separates storage by site origin, preventing cross-site tracking while maintaining functionality within individual websites.
Option a is incorrect because cookie sharing enables tracking by allowing sites to read cookies from other domains.
Option c is incorrect because cookie duplication creates copies but doesn't prevent cross-site access enabling user tracking.
Option d is incorrect because cookie compression reduces storage size but doesn't address privacy concerns about cross-site tracking.