Operating Systems practice questions

Randomly replace some recommendations with popular items.

Why wrong: Random replacements can hurt accuracy and do not guarantee diversity.

Use only popularity-based recommendations.

Why wrong: Popularity recommendations are uniform and reduce diversity.

Increase the number of recommendations and use collaborative filtering.

Why wrong: Collaborative filtering may already be in use; more recommendations may not improve diversity.

Supervised learning

Why wrong: Supervised learning requires labeled training data; recommendation systems often use unlabeled data.

Semi-supervised learning

Why wrong: Semi-supervised learning uses some labels; recommendation systems rarely have labels.

Transfer learning

Why wrong: Transfer learning is for reusing a model on a different task, not for building a new recommendation system from scratch.

Maximizing cost reduction for the hospital

Why wrong: Cost reduction is not an ethical priority.

Replacing human judgment entirely with AI

Why wrong: AI is meant to assist, not replace.

Using open-source models to reduce licensing costs

Why wrong: Licensing is not an ethical concern.

It typically has lower inference latency.

Why wrong: Latency depends on implementation, not architecture choice.

It handles novel scenarios better due to joint training.

Why wrong: End-to-end learning may generalize differently, but modular allows targeted improvements.

It requires less engineering effort.

Why wrong: Modular systems often require more engineering for interfaces.

Transparency

Why wrong: Transparency involves openness about model decisions, not addressing disparate impact.

Privacy

Why wrong: Privacy focuses on data protection, not group-level discrimination.

Accountability

Why wrong: Accountability is about assigning responsibility, not preventing bias itself.

Using open-source models only

Why wrong: Open-source does not ensure GDPR compliance.

Regular vulnerability scans

Why wrong: Scans address security, not privacy compliance.

Hiring more data scientists

Why wrong: Hiring does not automatically result in compliance.

Regular performance benchmarks

Why wrong: Not a specific requirement.

Open-source licensing

Why wrong: Not a requirement.

Mandatory use of cloud

Why wrong: Not mandated.

Data storage costs

Why wrong: Data storage is an operational cost, not an ethical concern.

Job displacement

Why wrong: Job displacement is a socioeconomic impact, not directly an ethical concern of automated decisions.

Environmental impact

Why wrong: Environmental impact is a separate sustainability issue.

Accountability

Why wrong: Accountability is about who is responsible, but the primary violation is bias.

Transparency

Why wrong: Transparency is about explainability, not directly addressing discrimination.

Privacy

Why wrong: Privacy relates to data protection, not bias.

SQL injection

Why wrong: Traditional database attack, not AI-specific.

Cross-site scripting

Why wrong: Web attack, not AI-specific.

Buffer overflow

Why wrong: Software vulnerability, not AI-specific.

Implementing differential privacy

Why wrong: Differential privacy protects privacy, not provides explanations.

Conducting bias audits

Why wrong: Bias audits detect unfairness but do not provide per-decision explanations.

Establishing an AI ethics board

Why wrong: An ethics board provides oversight but does not directly produce explanations.

Use an ensemble of models and majority voting

Why wrong: Ensembles do not inherently protect against adversarial attacks.

Increase the model's sensitivity to input changes

Why wrong: Higher sensitivity increases vulnerability to perturbations.

Reduce model complexity through pruning

Why wrong: Pruning may inadvertently remove robustness.

Obtain certification from an ISO 27001 auditor

Why wrong: ISO 27001 is for information security, not specifically for AI systems.

Publish the model's source code publicly

Why wrong: Open-source is not a requirement; proprietary systems are allowed.

Register the AI system with the national data protection authority

Why wrong: Registration is not specifically required for all high-risk systems; conformity assessment is.

Freeze the model to prevent any changes

Why wrong: Freezing the model would ignore drift and degrade performance.

Roll back to a previous model version if performance degrades

Why wrong: Rollback is a reactive measure, not a proactive management strategy.

Manually review all model predictions

Why wrong: Manual review is not scalable or efficient for drift management.

Remove all demographic features from the model.

Why wrong: Removing features may not eliminate bias if other correlated features exist.

Use a different algorithm that is inherently unbiased.

Why wrong: No algorithm is inherently unbiased; bias depends on data and deployment context.

Hire more diverse data scientists.

Why wrong: Diversity in team is beneficial but does not directly fix model bias.

Cross-validation

Why wrong: Cross-validation is a method to evaluate models, not a cause of bias.

High regularization

Why wrong: Regularization reduces variance, not bias.

Cloud-based deployment for scalability.

Why wrong: Deployment method is operational, not governance.

Model accuracy thresholds for production deployment.

Why wrong: Accuracy is a performance metric, not a governance component.

Use a different dataset from a similar hospital without checking demographics

Why wrong: Using a different dataset without validation may perpetuate or introduce new biases.

Manually increase the weight of age-related features in the model

Why wrong: Manual weighting can introduce new biases and lacks systematic validation.

Replace the neural network with a decision tree to simplify decision logic

Why wrong: Model simplification does not guarantee fairness and may reduce performance.

Association rule mining

Why wrong: Association rule mining finds frequent itemsets, not suited for sparse user-item matrices.

k-nearest neighbors

Why wrong: k-NN relies on similarity measures that may be inaccurate with sparse data.

Content-based filtering

Why wrong: Content-based filtering uses item attributes, not user-item interactions, and does not directly address matrix sparsity.

The cloud instance does not have enough memory

Why wrong: 32GB > 16GB, so memory should be sufficient unless other processes consume memory.

The training model is too large for the GPU memory

Why wrong: The error is from the system killing a process, not GPU-related.

The cloud instance ran out of disk space

Why wrong: The error is about memory, not disk space.