Wisdom of Committees: An Overlooked Approach To Faster and More Accurate Models
arXiv:2012.01988
Abstract
Committee-based models (ensembles or cascades) construct models by combining existing pre-trained ones. While ensembles and cascades are well-known techniques that were proposed before deep learning, they are not considered a core building block of deep model architectures and are rarely compared to in recent literature on developing efficient models. In this work, we go back to basics and conduct a comprehensive analysis of the efficiency of committee-based models. We find that even the most simplistic method for building committees from existing, independently pre-trained models can match or exceed the accuracy of state-of-the-art models while being drastically more efficient. These simple committee-based models also outperform sophisticated neural architecture search methods (e.g., BigNAS). These findings hold true for several tasks, including image classification, video classification, and semantic segmentation, and various architecture families, such as ViT, EfficientNet, ResNet, MobileNetV2, and X3D. Our results show that an EfficientNet cascade can achieve a 5.4x speedup over B7 and a ViT cascade can achieve a 2.3x speedup over ViT-L-384 while being equally accurate.
ICLR 2022
References in corpus (8)
- Rethinking Atrous Convolution for Semantic Image Segmentation
- On Calibration of Modern Neural Networks
- Deep Ensembles: A Loss Landscape Perspective
- Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
- BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning
- Dynamic Neural Networks: A Survey
- When Ensembling Smaller Models is More Efficient than Single Large Models
- Learnable Embedding Space for Efficient Neural Architecture Compression