Corpus ID: 173990166

Discovering Neural Wirings

  title={Discovering Neural Wirings},
  author={Mitchell Wortsman and Ali Farhadi and M. Rastegari},
  • Mitchell Wortsman, Ali Farhadi, M. Rastegari
  • Published in NeurIPS 2019
  • Computer Science
  • The success of neural networks has driven a shift in focus from feature engineering to architecture engineering. However, successful networks today are constructed using a small and manually defined set of building blocks. Even in methods of neural architecture search (NAS) the network connectivity patterns are largely constrained. In this work we propose a method for discovering neural wirings. We relax the typical notion of layers and instead enable channels to form connections independent of… CONTINUE READING
    28 Citations

    Figures, Tables, and Topics from this paper.

    Revisiting Neural Architecture Search
    Understanding the wiring evolution in differentiable neural architecture search
    Graph Structure of Neural Networks
    • 4
    • PDF
    Neural networks adapting to datasets: learning network size and topology
    How Does Topology of Neural Architectures Impact Gradient Propagation and Model Performance
    What’s Hidden in a Randomly Weighted Neural Network?
    • 41
    • PDF
    Rigging the Lottery: Making All Tickets Winners
    • 35
    • PDF
    Explicitly Learning Topology for Differentiable Neural Architecture Search
    Neural Architecture Generator Optimization


    Exploring Randomly Wired Neural Networks for Image Recognition
    • 146
    • Highly Influential
    • PDF
    Neural Architecture Search with Reinforcement Learning
    • 2,066
    • PDF
    Learning Implicitly Recurrent CNNs Through Parameter Sharing
    • 16
    • PDF
    Densely Connected Convolutional Networks
    • 10,506
    • PDF
    Learning Sparse Networks Using Targeted Dropout
    • 22
    • PDF
    Learning Multiple Layers of Features from Tiny Images
    • 9,378
    • Highly Influential
    • PDF
    The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
    • 522
    • PDF
    Learning Sparse Neural Networks through L0 Regularization
    • 331
    • PDF