Ensuring Reliability via Hyperparameter Selection: Review and Advances
This work addresses reliability in AI deployment, particularly for foundational models, but is incremental as it builds on existing frameworks.
The paper tackles hyperparameter selection for AI models by framing it as a multiple hypothesis testing problem, reviewing the Learn-Then-Test framework and extending it with new risk measures, multi-objective optimization, and prior knowledge integration, with applications in communication systems.
Hyperparameter selection is a critical step in the deployment of artificial intelligence (AI) models, particularly in the current era of foundational, pre-trained, models. By framing hyperparameter selection as a multiple hypothesis testing problem, recent research has shown that it is possible to provide statistical guarantees on population risk measures attained by the selected hyperparameter. This paper reviews the Learn-Then-Test (LTT) framework, which formalizes this approach, and explores several extensions tailored to engineering-relevant scenarios. These extensions encompass different risk measures and statistical guarantees, multi-objective optimization, the incorporation of prior knowledge and dependency structures into the hyperparameter selection process, as well as adaptivity. The paper also includes illustrative applications for communication systems.