The correct answer is D) AutoML efficiently distributes hyperparameter tuning trials across multiple worker nodes. Here's why:

• Distributed Hyperparameter Tuning: AutoML excels in parallelizing hyperparameter tuning, utilizing Spark's distributed architecture for enhanced efficiency.
• Trial Execution: It smartly allocates hyperparameter trials across various worker nodes, allowing for the concurrent exploration of different model configurations.
• Accelerated Optimization: This approach significantly quickens the model tuning process, particularly beneficial for large datasets and complex models.

Clarifications on Other Options:

• A) Incorrect: AutoML is capable of processing large datasets by partitioning them for distributed processing.
• B) Incorrect: Beyond training models, AutoML also assesses their performance using multiple metrics to identify the top-performing model.
• C) Incorrect: For optimal performance and to prevent potential conflicts, AutoML typically necessitates a cluster in exclusive mode.

Key Insights: AutoML's ability to distribute hyperparameter tuning tasks is a pivotal feature that enhances its efficiency and effectiveness in discovering high-performing models. Grasping this functionality is essential for leveraging AutoML to optimize machine learning workflows in distributed settings.