Speaker
Description
As machine learning models become increasingly integrated into various applications, the need for resource-aware deployment strategies becomes paramount. One promising approach for optimizing resource consumption is rejection ensembles. Rejection ensembles combine a small model deployed to an edge device with a large model deployed in the cloud, with a rejector tasked to determine the most suitable model for a given input. Due to its novelty, existing research predominantly focuses on ad-hoc ensemble design, lacking a thorough understanding of rejector optimization and deployment strategies. In this talk, we focus on this research gap by presenting a theoretical investigation into rejection ensembles and proposing a novel algorithm for training and deploying rejectors based on these novel insights. We give precise conditions of when a good rejector can improve the ensemble's overall performance beyond the big model's performance, and when a bad rejector can make the ensemble worse than the small model. Second, we show that even the perfect rejector can overuse its budget for using the big model during deployment. Based on these insights, we propose to ignore any budget constraints during training but introduce additional safeguards during deployment. Experimental evaluation on 8 different datasets from various domains demonstrates the efficacy of our novel rejection ensemble outperforming existing approaches. Moreover, compared to standalone large model inference, we highlight the energy efficiency gains during deployment on a Nvidia Jetson AGX board.
Note: This work has been published and presented at the ECML-PKDD 2024.
