AutoML in Heavily Constrained Applications
This work addresses the need for more flexible and efficient AutoML in constrained applications, representing an incremental improvement over existing systems.
The paper tackles the problem of AutoML systems lacking adaptability to specific use cases and user-defined constraints, proposing CAML, which uses meta-learning to automatically adapt AutoML parameters and achieve high predictive performance while satisfying constraints.
Optimizing a machine learning pipeline for a task at hand requires careful configuration of various hyperparameters, typically supported by an AutoML system that optimizes the hyperparameters for the given training dataset. Yet, depending on the AutoML system's own second-order meta-configuration, the performance of the AutoML process can vary significantly. Current AutoML systems cannot automatically adapt their own configuration to a specific use case. Further, they cannot compile user-defined application constraints on the effectiveness and efficiency of the pipeline and its generation. In this paper, we propose CAML, which uses meta-learning to automatically adapt its own AutoML parameters, such as the search strategy, the validation strategy, and the search space, for a task at hand. The dynamic AutoML strategy of CAML takes user-defined constraints into account and obtains constraint-satisfying pipelines with high predictive performance.