Efficient NAS via Parameter Sharing

This paper proposes a novel approach to Efficient Neural Architecture Search (NAS) using parameter sharing. The authors introduce a new search space and an efficient algorithm to explore it, aiming to reduce the computational cost of NAS.

  • Summary

    • The paper introduces a parameter-sharing-based NAS approach.
    • It proposes a new search space and an efficient algorithm.
    • The method aims to reduce the computational cost of NAS.
  • Key Points

    • Parameter Sharing: The proposed method uses parameter sharing to reduce the number of parameters in the search space.
    • Efficient Search Space: A new search space is introduced, which is more efficient to explore.
    • Algorithm: An efficient algorithm is proposed to explore the new search space.
  • Applications

    • The proposed approach can be applied to various domains, such as computer vision, natural language processing, and speech recognition.
  • Further Reading

  • Image

    Parameter Sharing