שאלה | 13. [Meta-learning] Which of the following statements best characterizes the Model Agnostic Meta-Learning (MAML) algorithm:

13. [Meta-learning] Which of the following statements best characterizes the Model Agnostic Meta-Learning (MAML) algorithm:

		a) MAML concurrently trains separate models for each task, optimizing them jointly to ensure a task-specific performance
		b) MAML initializes a single model with random weights and trains it on all tasks simultaneously, adjusting its parameters to facilitate rapid adaptation to new tasks
		c) MAML employs reinforcement learning to sequentially train a single model on each task, leveraging task-specific rewards to guide the learning process
		d) MAML trains by iteratively updating a single model’s parameters based on its ability to quickly adapt to new tasks, ensur-ing effective generalization across task domains