Hyperparameter Learning via Distributional Transfer
This work addresses hyperparameter tuning for machine learning practitioners, offering an incremental improvement over existing Bayesian optimization methods.
The paper tackles the problem of slow hyperparameter optimization by transferring information from prior tasks using learned dataset representations, resulting in a joint Gaussian process model that achieves faster convergence, sometimes requiring only a few objective evaluations.
Bayesian optimisation is a popular technique for hyperparameter learning but typically requires initial exploration even in cases where similar prior tasks have been solved. We propose to transfer information across tasks using learnt representations of training datasets used in those tasks. This results in a joint Gaussian process model on hyperparameters and data representations. Representations make use of the framework of distribution embeddings into reproducing kernel Hilbert spaces. The developed method has a faster convergence compared to existing baselines, in some cases requiring only a few evaluations of the target objective.