Ensembles of Locally Independent Prediction Models
This addresses a practical issue for machine learning practitioners using ensembles in non-stationary environments, but it is incremental as it builds on existing ensemble methods with a new diversity focus.
The paper tackles the problem that traditional ensemble diversity metrics based on training set predictions fail under covariate shift, harming generalization. It introduces a new diversity metric and training method for ensembles that extrapolate differently locally, improving generalization and diversity in synthetic and real-world tasks, especially under data limits and covariate shift.
Ensembles depend on diversity for improved performance. Many ensemble training methods, therefore, attempt to optimize for diversity, which they almost always define in terms of differences in training set predictions. In this paper, however, we demonstrate the diversity of predictions on the training set does not necessarily imply diversity under mild covariate shift, which can harm generalization in practical settings. To address this issue, we introduce a new diversity metric and associated method of training ensembles of models that extrapolate differently on local patches of the data manifold. Across a variety of synthetic and real-world tasks, we find that our method improves generalization and diversity in qualitatively novel ways, especially under data limits and covariate shift.