Contrastive Embeddings for Neural Architectures

The performance of algorithms for neural architecture search strongly depends on the parametrization of the search space. We use contrastive learning to identify networks across different initializations based on their data Jacobians, and automatically produce the first architecture embeddings independent from the parametrization of the search space. Using our contrastive embeddings, we show that traditional black-box optimization algorithms, without modification, can reach state-of-the-art performance in Neural Architecture Search. As our method provides a unified embedding space, we perform for the first time transfer learning between search spaces. Finally, we show the evolution of embeddings during training, motivating future studies into using embeddings at different training stages to gain a deeper understanding of the networks in a search space.

[1]  Kaiming He,et al.  Momentum Contrast for Unsupervised Visual Representation Learning , 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Amos Storkey,et al.  Neural Architecture Search without Training , 2021, ICML.

[3]  Razvan Pascanu,et al.  Sobolev Training for Neural Networks , 2017, NIPS.

[4]  Vijay Vasudevan,et al.  Learning Transferable Architectures for Scalable Image Recognition , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[5]  Alok Aggarwal,et al.  Regularized Evolution for Image Classifier Architecture Search , 2018, AAAI.

[6]  Yoshua Bengio,et al.  Algorithms for Hyper-Parameter Optimization , 2011, NIPS.

[7]  Geoffrey E. Hinton,et al.  Visualizing Data using t-SNE , 2008 .

[8]  Mi Zhang,et al.  Does Unsupervised Architecture Representation Learning Help Neural Architecture Search? , 2020, NeurIPS.

[9]  Haihong Hu,et al.  Self-Supervised Representation Learning for Evolutionary Neural Architecture Search , 2020, IEEE Computational Intelligence Magazine.

[10]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[11]  Yi Yang,et al.  NAS-Bench-102: Extending the Scope of Reproducible Neural Architecture Search , 2020 .

[12]  Julien Mairal,et al.  Unsupervised Learning of Visual Features by Contrasting Cluster Assignments , 2020, NeurIPS.

[13]  Yuandong Tian,et al.  Sample-Efficient Neural Architecture Search by Learning Action Space , 2019, ArXiv.

[14]  Geoffrey E. Hinton,et al.  A Simple Framework for Contrastive Learning of Visual Representations , 2020, ICML.

[15]  Tie-Yan Liu,et al.  LightGBM: A Highly Efficient Gradient Boosting Decision Tree , 2017, NIPS.

[16]  Michal Valko,et al.  Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning , 2020, NeurIPS.

[17]  François Fleuret,et al.  Knowledge Transfer with Jacobian Matching , 2018, ICML.

[18]  Arthur Jacot,et al.  Neural tangent kernel: convergence and generalization in neural networks (invited paper) , 2018, NeurIPS.

[19]  Jeff A. Bilmes,et al.  Analysis of Deep Neural Networks with Extended Data Jacobian Matrix , 2016, ICML.