Answer-first summary for fast verification
Answer: Switch the iterative optimization algorithm to one better suited for the problem.
If no consistent improvement in model accuracy is observed during hyperparameter tuning, the current optimization algorithm may not be appropriate for the problem. Switching to a different algorithm could more effectively explore the hyperparameter space, potentially leading to better model accuracy. Adjustments to the number of compute nodes primarily affect computation speed rather than the effectiveness of the hyperparameter search.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
No comments yet.
A data scientist is performing hyperparameter optimization using an iterative optimization algorithm, with each unique set of hyperparameters trained on a separate compute node. Despite conducting eight evaluations across eight nodes, they notice no consistent improvement in model accuracy. What changes could potentially enhance the model's accuracy during this tuning process?
A
Switch the iterative optimization algorithm to one better suited for the problem.
B
Increase both the number of compute nodes and evaluations significantly.
C
Reduce the number of compute nodes to half or less than half of the evaluations.
D
Double or more than double the number of compute nodes relative to evaluations.
E
Significantly decrease both the number of compute nodes and evaluations.