Efficient Stochastic Inference of Bitwise Deep Neural Networks

@article{Vogel2016EfficientSI,
  title={Efficient Stochastic Inference of Bitwise Deep Neural Networks},
  author={Sebastian Vogel and Christoph Schorn and Andre Guntoro and Gerd Ascheid},
  journal={CoRR},
  year={2016},
  volume={abs/1611.06539}
}
Recently published methods enable training of bitwise neural networks which allow reduced representation of down to a single bit per weight. We present a method that exploits ensemble decisions based on multiple stochastically sampled network models to increase performance figures of bitwise neural networks in terms of classification accuracy at inference. Our experiments with the CIFAR-10 and GTSRB datasets show that the performance of such network ensembles surpasses the performance of the… CONTINUE READING
  • GitHub repos referencing this paper

  • Presentations referencing similar topics