Corpus ID: 53733261

Understanding the Importance of Single Directions via Representative Substitution

@article{Chen2018UnderstandingTI,
  title={Understanding the Importance of Single Directions via Representative Substitution},
  author={L. Chen and Hailun Ding and Qi Li and Z. Li and Jian Peng and Haifeng Li},
  journal={ArXiv},
  year={2018},
  volume={abs/1811.11053}
}
  • L. Chen, Hailun Ding, +3 authors Haifeng Li
  • Published 2018
  • Mathematics, Computer Science
  • ArXiv
  • Understanding the internal representations of deep neural networks (DNNs) is crucal to explain their behavior. The interpretation of individual units, which are neurons in MLPs or convolution kernels in convolutional networks, has been paid much attention given their fundamental role. However, recent research (Morcos et al. 2018) presented a counterintuitive phenomenon, which suggests that an individual unit with high class selectivity, called interpretable units, has poor contributions to… CONTINUE READING

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 11 REFERENCES
    On the Expressive Power of Deep Neural Networks
    308