Anytime Inference with Distilled Hierarchical Neural Ensembles
arXiv:2003.01474
Abstract
Inference in deep neural networks can be computationally expensive, and networks capable of anytime inference are important in mscenarios where the amount of compute or quantity of input data varies over time. In such networks the inference process can interrupted to provide a result faster, or continued to obtain a more accurate result. We propose Hierarchical Neural Ensembles (HNE), a novel framework to embed an ensemble of multiple networks in a hierarchical tree structure, sharing intermediate layers. In HNE we control the complexity of inference on-the-fly by evaluating more or less models in the ensemble. Our second contribution is a novel hierarchical distillation method to boost the prediction accuracy of small ensembles. This approach leverages the nested structure of our ensembles, to optimally allocate accuracy and diversity across the individual models. Our experiments show that, compared to previous anytime inference models, HNE provides state-of-the-art accuracy-computate trade-offs on the CIFAR-10/100 and ImageNet datasets.
References in corpus (18)
- MobileNetV2: Inverted Residuals and Linear Bottlenecks
- SGDR: Stochastic Gradient Descent with Warm Restarts
- Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
- Learning Efficient Convolutional Networks through Network Slimming
- Multi-Scale Dense Networks for Resource Efficient Image Classification
- Scaling description of generalization with number of parameters in deep learning
- Large scale distributed neural network training through online distillation
- Hydra: an Ensemble of Convolutional Neural Networks for Geospatial Land Classification
- Adaptive Neural Networks for Efficient Inference
- Once-for-All: Train One Network and Specialize it for Efficient Deployment
- CondenseNet: An Efficient DenseNet using Learned Group Convolutions
- FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search
- SkipNet: Learning Dynamic Routing in Convolutional Networks
- BlockDrop: Dynamic Inference Paths in Residual Networks
- BigNAS: Scaling Up Neural Architecture Search with Big Single-Stage Models
- Resolution Adaptive Networks for Efficient Inference
- Improved Techniques for Training Adaptive Deep Networks
- Universally Slimmable Networks and Improved Training Techniques