AxTrain: Hardware-Oriented Neural Network Training for Approximate Inference

@article{He2018AxTrainHN,
  title={AxTrain: Hardware-Oriented Neural Network Training for Approximate Inference},
  author={Xin He and Liu Ke and W. Lu and Guihai Yan and Xuan Zhang},
  journal={Proceedings of the International Symposium on Low Power Electronics and Design},
  year={2018}
}
  • Xin He, Liu Ke, +2 authors Xuan Zhang
  • Published 2018
  • Computer Science, Engineering, Mathematics
  • Proceedings of the International Symposium on Low Power Electronics and Design
  • The intrinsic error tolerance of neural network (NN) makes approximate computing a promising technique to improve the energy efficiency of NN inference. [...] Key Method Specifically, AxTrain leverages the synergy between two orthogonal methods---one actively searches for a network parameters distribution with high error tolerance, and the other passively learns resilient weights by numerically incorporating the noise distributions of the approximate hardware in the forward pass during the training phase…Expand Abstract
    17 Citations
    Full Approximation of Deep Neural Networks through Efficient Optimization
    AxDNN: towards the cross-layer design of approximate DNNs
    • 4
    Improving approximate neural networks for perception tasks through specialized optimization
    CANN: Curable Approximations for High-Performance Deep Neural Network Accelerators
    • 3
    HeadStart: Enforcing Optimal Inceptions in Pruning Deep Neural Networks for Efficient Inference on GPGPUs
    • 1
    A Quantitative Exploration of Collaborative Pruning and Approximation Computing Towards Energy Efficient Neural Networks
    • 1
    ProxSim: GPU-based Simulation Framework for Cross-Layer Approximate DNN Optimization
    • 4
    • Highly Influenced
    • PDF
    NeuADC: Neural Network-Inspired RRAM-Based Synthesizable Analog-to-Digital Conversion with Reconfigurable Quantization Support
    • 3
    Stochasticity and Robustness in Spiking Neural Networks
    • 2
    • PDF

    References

    Near-Threshold Computing: Reclaiming Moore's Law Through Energy Efficient Integrated Circuits
    • 731
    • Highly Influential
    • PDF