Answer: (b)

• • Explanation: Deep learning models can be complex, making it difficult to understand how they arrive at decisions (b). They often require significant computational power (c) and can handle various data formats with proper training. Ease of use is improving, though challenges remain (a).

Answer: (c)

• • Explanation: Biases in training data can be reflected in algorithms, leading to unfair outcomes (c). Algorithmic bias doesn't improve training or accuracy (a) or promote fairness (b). While it might not directly impact data collection (d), addressing bias in existing data is crucial.

Answer: (d)

• • Explanation: XAI techniques aim to make deep learning models more understandable (d), fostering trust and accountability. While it might not directly impact accuracy or efficiency (a), it doesn't necessarily reduce human oversight (b) or training speed (c).

Answer: (b)

• • Explanation: Deep learning excels at identifying patterns in large datasets, unlike traditional AI focused on symbolic reasoning (a) or faster computations (c). While advancements are being made in human-like interactions (d), deep learning's core strength lies in pattern recognition from data.

Answer: (b)

• • Explanation: PPPs bring together government, academia, and industry (b), fostering collaboration for responsible development. While they can reduce costs (c), the primary focus is collaboration, not commercial advantage (a) or prioritizing economic benefits over ethics (d).