Boosted Generative Models
This work addresses the challenge of boosting generative models for researchers and practitioners, offering a meta-algorithmic approach that is incremental in nature.
The paper tackles the problem of improving generative models by introducing an unsupervised boosting framework that sequentially trains models to correct earlier mistakes, achieving enhanced performance in density estimation, classification, and sample generation on benchmark datasets.
We propose a novel approach for using unsupervised boosting to create an ensemble of generative models, where models are trained in sequence to correct earlier mistakes. Our meta-algorithmic framework can leverage any existing base learner that permits likelihood evaluation, including recent deep expressive models. Further, our approach allows the ensemble to include discriminative models trained to distinguish real data from model-generated data. We show theoretical conditions under which incorporating a new model in the ensemble will improve the fit and empirically demonstrate the effectiveness of our black-box boosting algorithms on density estimation, classification, and sample generation on benchmark datasets for a wide range of generative models.