D429 Introduction to AI for Computer Scientists - Set 2 - Part 3

Correct Answer: Action sequence recognition

Option b is correct because recognizing action sequences can identify unsafe behaviors that only become apparent through temporal movement patterns analyzed.
Option a is incorrect because single frames miss the temporal context needed to identify unsafe actions that unfold over time periods.
Option c is incorrect because color detection alone cannot identify unsafe behaviors that depend on worker movements and equipment interactions dynamically.
Option d is incorrect because static comparisons miss the dynamic nature of safety violations that involve movement and temporal sequences.

Correct Answer: Emotional trajectory modeling

Option b is correct because trajectory modeling tracks emotional intensity changes over time, identifying escalation patterns before complaints formalize completely.
Option a is incorrect because word counting misses the emotional dynamics and contextual buildup that characterizes conversation escalation toward complaints progressively.
Option c is incorrect because analyzing only final messages misses the escalation process, making prevention of formal complaints impossible effectively.
Option d is incorrect because random sampling fails to capture the continuous progression of emotional escalation in customer service conversations.

Correct Answer: Causal inference testing

Option b is correct because causal inference methods can distinguish genuine cause-effect relationships from mere statistical correlations in market data.
Option a is incorrect because correlation matrices show relationships but cannot determine whether one factor actually causes another versus coincidental correlation.
Option c is incorrect because assuming random relationships ignores both correlations and causal connections that are crucial for investment decisions.
Option d is incorrect because trending analysis identifies patterns over time but cannot establish whether relationships are causal or coincidental.

Correct Answer: Adaptive clinical learning

Option b is correct because adaptive learning can modify treatment recommendations based on observed patient responses, optimizing personalized care over time.
Option a is incorrect because static protocols cannot adapt to individual patient variations or learn from treatment outcomes to improve recommendations.
Option c is incorrect because fixed treatment paths ignore individual patient responses and cannot optimize based on observed treatment effectiveness patterns.
Option d is incorrect because random selection provides no mechanism for learning which treatments work best for specific patient characteristics.

Correct Answer: Anomaly behavior detection

Option b is correct because anomaly detection can identify unusual behaviors indicating new attacks without requiring prior examples or signatures.
Option a is incorrect because signature matching only detects known attacks and cannot identify zero-day exploits lacking established signature patterns.
Option c is incorrect because known threat databases by definition don't contain information about zero-day attacks that haven't been discovered yet.
Option d is incorrect because historical pattern matching relies on past attacks and cannot detect novel attack methods never seen before.

Correct Answer: Constrained optimization approach

Option b is correct because constrained optimization can minimize pesticide use while maintaining effectiveness constraints for adequate crop protection levels.
Option a is incorrect because maximum application wastes chemicals, harms the environment, and may damage crops through excessive pesticide exposure unnecessarily.
Option c is incorrect because random patterns lead to uneven coverage, leaving some areas unprotected while over-treating others with excessive chemicals.
Option d is incorrect because fixed schedules ignore actual pest presence and weather conditions, leading to ineffective or wasteful pesticide application.

Correct Answer: Semantic similarity analysis

Option b is correct because semantic analysis can identify conceptual similarities even when different terminology describes the same technical innovations effectively.
Option a is incorrect because exact matching misses patent infringements that describe the same concepts using different technical terms or phrasing.
Option c is incorrect because word frequency doesn't capture the meaning and relationships that determine whether technical concepts infringe existing patents.
Option d is incorrect because format comparison ignores the actual content and concepts that determine patent infringement in technical documents.

Correct Answer: Dynamic priority adjustment

Option b is correct because dynamic adjustment can temporarily prioritize emergency routes while minimizing overall traffic disruption across the network.
Option a is incorrect because optimizing only for emergencies would create severe congestion for normal traffic, potentially blocking other emergency routes.
Option c is incorrect because fixed lane allocation reduces road capacity for normal traffic even when no emergencies are occurring currently.
Option d is incorrect because equal treatment delays emergency response, potentially costing lives when rapid medical or fire response is critical.

Correct Answer: Citation requirement enforcement

Option b is correct because requiring citations ensures generated content properly attributes sources, maintaining academic integrity standards throughout the writing.
Option a is incorrect because unrestricted generation could produce text that appears original but actually reproduces existing work without proper attribution.
Option c is incorrect because allowing copy-paste functionality enables direct plagiarism without the transformative work required for academic integrity compliance.
Option d is incorrect because hiding sources while using their content constitutes plagiarism, violating fundamental academic integrity principles and standards.

Correct Answer: Pattern-based risk detection

Option b is correct because pattern detection can identify crisis indicators without storing or deeply analyzing individual personal communications unnecessarily.
Option a is incorrect because complete scanning violates privacy by examining all messages regardless of risk indicators or public safety relevance.
Option c is incorrect because individual tracking creates privacy violations and surveillance concerns beyond what's necessary for crisis intervention purposes.
Option d is incorrect because unrestricted data mining violates privacy principles and likely breaches regulations protecting user communication data rights.