NCA-AIIO Practice Questions

A deep learning model achieves very high ROC-AUC but performs poorly in real-world deployment. Which is the MOST plausible explanation?

Which situation MOST strongly indicates data leakage?

A model generalizes poorly even with abundant training data. Which factor is MOST likely responsible?

Why do convolutional neural networks outperform fully connected networks in vision tasks?

Which property MOST explains why deep networks can represent complex functions efficiently?

In gradient-based optimization, feature scaling is crucial because it:

Which factor MOST increases variance in reinforcement learning policy gradient methods?

Which scenario BEST illustrates concept drift?

Why do attention-based models (Transformers) scale better than RNNs for long sequences?

Which metric is LEAST appropriate for evaluating recommendation ranking quality?

Which phenomenon explains performance degradation when adding many irrelevant features?

Why is explainability critical in safety-critical AI systems?

In transfer learning, when is negative transfer MOST likely?

Which assumption is violated when deployment data follows a different distribution than training data?

Which statement BEST characterizes Artificial General Intelligence (AGI)?