Journal
ENTROPY
Volume 24, Issue 1, Pages -Publisher
MDPI
DOI: 10.3390/e24010001
Keywords
branch neural networks; deep learning; deep neural networks; adaptive computation; fast inference
Categories
Ask authors/readers for more resources
This paper introduces a new approach to training deep neural networks with multiple intermediate auxiliary classifiers, which can reduce inference time while maintaining prediction accuracy. By using trained confidence scores, the multi-exit model is trained without hyper-parameters.
In this paper, we propose a new approach to train a deep neural network with multiple intermediate auxiliary classifiers, branching from it. These 'multi-exits' models can be used to reduce the inference time by performing early exit on the intermediate branches, if the confidence of the prediction is higher than a threshold. They rely on the assumption that not all the samples require the same amount of processing to yield a good prediction. In this paper, we propose a way to train jointly all the branches of a multi-exit model without hyper-parameters, by weighting the predictions from each branch with a trained confidence score. Each confidence score is an approximation of the real one produced by the branch, and it is calculated and regularized while training the rest of the model. We evaluate our proposal on a set of image classification benchmarks, using different neural models and early-exit stopping criteria.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available