Corpus ID: 52180104

Probabilistic Binary Neural Networks

  title={Probabilistic Binary Neural Networks},
  author={Jorn W. T. Peters and M. Welling},
  • Jorn W. T. Peters, M. Welling
  • Published 2018
  • Computer Science, Mathematics
  • ArXiv
  • Low bit-width weights and activations are an effective way of combating the increasing need for both memory and compute power of Deep Neural Networks. In this work, we present a probabilistic training method for Neural Network with both binary weights and activations, called BLRNet. By embracing stochasticity during training, we circumvent the need to approximate the gradient of non-differentiable functions such as sign(), while still obtaining a fully Binary Neural Network at test time… CONTINUE READING
    Relaxed Quantization for Discretized Neural Networks
    • 50
    • Open Access
    Data-Free Quantization Through Weight Equalization and Bias Correction
    • 49
    • Open Access
    Adv-BNN: Improved Adversarial Defense through Robust Bayesian Neural Network
    • 45
    • Open Access
    Bit Efficient Quantization for Deep Neural Networks
    • 7
    • Open Access
    Latent Weights Do Not Exist: Rethinking Binarized Neural Network Optimization
    • 16
    • Open Access
    Resource-Efficient Neural Networks for Embedded Systems
    • 3
    • Open Access
    BAMSProd: A Step towards Generalizing the Adaptive Optimization Methods to Deep Binary Model
    • 1
    • Open Access
    Signal propagation in continuous approximations of binary neural networks
    Progressive Stochastic Binarization of Deep Networks
    Reintroducing Straight-Through Estimators as Principled Methods for Stochastic Binary Networks


    Publications referenced by this paper.
    Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
    • 18,716
    • Highly Influential
    • Open Access
    Adam: A Method for Stochastic Optimization
    • 49,112
    • Open Access
    Binarized Neural Networks
    • 761
    • Highly Influential
    • Open Access
    Understanding the difficulty of training deep feedforward neural networks
    • 8,428
    • Open Access
    Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding
    • 3,650
    • Open Access
    Towards Accurate Binary Convolutional Neural Network
    • 246
    • Open Access
    Wide Residual Networks
    • 2,286
    • Open Access
    Auto-Encoding Variational Bayes
    • 9,051
    • Open Access
    Learning Multiple Layers of Features from Tiny Images
    • 9,041
    • Highly Influential
    • Open Access
    Stochastic Pooling for Regularization of Deep Convolutional Neural Networks
    • 681
    • Open Access