Corpus ID: 210700033

MeliusNet: Can Binary Neural Networks Achieve MobileNet-level Accuracy?

@article{Bethge2020MeliusNetCB,
  title={MeliusNet: Can Binary Neural Networks Achieve MobileNet-level Accuracy?},
  author={Joseph Bethge and C. Bartz and Haojin Yang and Y. Chen and C. Meinel},
  journal={ArXiv},
  year={2020},
  volume={abs/2001.05936}
}
  • Joseph Bethge, C. Bartz, +2 authors C. Meinel
  • Published 2020
  • Computer Science, Mathematics
  • ArXiv
  • Binary Neural Networks (BNNs) are neural networks which use binary weights and activations instead of the typical 32-bit floating point values. They have reduced model sizes and allow for efficient inference on mobile or embedded devices with limited power and computational resources. However, the binarization of weights and activations leads to feature maps of lower quality and lower capacity and thus a drop in accuracy compared to traditional networks. Previous work has increased the number… CONTINUE READING
    17 Citations

    Figures, Tables, and Topics from this paper

    FracBNN: Accurate and FPGA-Efficient Binary Neural Networks with Fractional Activations
    • Highly Influenced
    • PDF
    Larq Compute Engine: Design, Benchmark, and Deploy State-of-the-Art Binarized Neural Networks
    • Highly Influenced
    • PDF
    Rotated Binary Neural Network
    • 3
    • PDF
    SiMaN: Sign-to-Magnitude Network Binarization
    • PDF
    OPTIMIZING QUANTIZED NEURAL NETWORKS IN A WEAK CURVATURE MANIFOLD
    • 2020

    References

    SHOWING 1-10 OF 49 REFERENCES
    Training Competitive Binary Neural Networks from Scratch
    • 23
    • PDF
    Binarized Neural Networks
    • 896
    • PDF
    Trained Ternary Quantization
    • 621
    • PDF
    XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks
    • 1,683
    • Highly Influential
    • PDF
    BinaryConnect: Training Deep Neural Networks with binary weights during propagations
    • 1,652
    • PDF
    Towards Accurate Binary Convolutional Neural Network
    • 293
    • Highly Influential
    • PDF
    BinaryDenseNet: Developing an Architecture for Binary Neural Networks
    • 11
    • PDF
    Binary Ensemble Neural Network: More Bits per Network or More Networks per Bit?
    • Shilin Zhu, Xin Dong, H. Su
    • Computer Science, Mathematics
    • 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
    • 2019
    • 62
    • Highly Influential
    • PDF
    XNOR-Net++: Improved binary neural networks
    • 35
    • PDF
    Learning to Quantize Deep Networks by Optimizing Quantization Intervals With Task Loss
    • 116
    • PDF