著者: Sanjeev Arora , Aditya Bhaskara , Rong Ge , Tengyu Ma - International Conference on Machine Learning, ICML 2014 被引用: 342
We give algorithms with provable guarantees that learn a class of deep nets in the generative model view popularized by Hinton and others. Our generative model is an $n$ node multilayer neural net that has degree at most $n^γ$ for some $γ<1$ and each edge has a random edge weight in $[-1,1]$. Our algorithm learns {\em almost all} networks in this class with polynomial running time. The sample complexity is quadratic or cubic depending upon the details of the model. The algorithm uses layerwise learning. It is based upon a novel idea of observing correlations among features and using these to infer the underlying edge structure via a global graph recovery procedure. The analysis of the algorithm reveals interesting structure of neural networks with random edge weights.
✨ ログイン状態を確認しています… PDF 被引用 BibTeX を表示 BibTeX を閉じる BibTeX を表示 引用
The Organization of Behavior; A… The Organization of Behavior; A Neuropsychological Theory Compressed sensing Compressed sensing Expander graphs and their applications Expander graphs and their applications Extracting and composing robust features with… Extracting and composing robust features with denoising autoencoders Combining geometry and combinatorics: A unifie… Combining geometry and combinatorics: A unified approach to sparse signal recovery Kernel Methods for Deep Learning Kernel Methods for Deep Learning Learning Deep Architectures for AI Learning Deep Architectures for AI On Random Weights and Unsupervised Feature… On Random Weights and Unsupervised Feature Learning ImageNet Classification with Deep Convolutional… ImageNet Classification with Deep Convolutional Neural Networks Learning mixtures of spherical gaussians… Learning mixtures of spherical gaussians: moment methods and spectral decompositions Unsupervised Feature Learning and Deep… Unsupervised Feature Learning and Deep Learning: A Review and New Perspectives New Algorithms for Learning Incoherent and… New Algorithms for Learning Incoherent and Overcomplete Dictionaries Sparse Matrix Factorization Sparse Matrix Factorization Going Deeper with Convolutions Going Deeper with Convolutions Generalization Bounds for Neural Networks… Generalization Bounds for Neural Networks through Tensor Factorization Provable Methods for Training Neural Network… Provable Methods for Training Neural Networks with Sparse Connectivity A Probabilistic Theory of Deep Learning A Probabilistic Theory of Deep Learning l 1 -regularized neural networks are improperly… l 1 -regularized neural networks are improperly learnable in polynomial time Recovery Guarantees for One-hidden-layer Neural… Recovery Guarantees for One-hidden-layer Neural Networks AdaNet: Adaptive Structural Learning of… AdaNet: Adaptive Structural Learning of Artificial Neural Networks Learning One-hidden-layer Neural… Learning One-hidden-layer Neural Networks with Landscape Design A mean field view of the landscape of two-layer… A mean field view of the landscape of two-layer neural networks On the Convergence Rate of Training Recurrent… On the Convergence Rate of Training Recurrent Neural Networks Learning Two Layer Rectified Neural… Learning Two Layer Rectified Neural Networks in Polynomial Time Provable Bounds for Learning Some Deep… Provable Bounds for Learning Some Deep Representations 過去の参考文献 中心の論文 この論文を引用する論文 古い 新しい ノードをクリックするとフォーカスを固定、空白をクリックすると本論文に戻ります。ホバーで一時的にプレビューできます。各ノードのページはタイトルから開けます。
