Theory

An Empirical Study on The Properties of Random Bases for Kernel Methods NeurIPS 2017 Axiomatic Attribution for Deep Networks ICML 2017 Convergence Analysis of Two-layer Neural Networks with ReLU Activation NeurIPS 2017 Deep Hyperspherical Learning NeurIPS 2017 Eigenvalue Decay Implies Polynomial-Time Learnability for Neural Networks NeurIPS 2017 Failures of Gradient-Based Deep Learning ICML 2017 Fixed-Rank Approximation of a Positive-Semidefinite Matrix from Streaming Data NeurIPS 2017 Global Optimality in Neural Network Training CVPR 2017 Gradient Descent Can Take Exponential Time to Escape Saddle Points NeurIPS 2017 How regularization affects the critical points in linear networks NeurIPS 2017 Implicit Regularization in Matrix Factorization NeurIPS 2017 Improved Graph Laplacian via Geometric Self-Consistency NeurIPS 2017 Invariance and Stability of Deep Convolutional Representations NeurIPS 2017 Limits of End-to-End Learning ACML 2017 Mean Field Residual Networks: On the Edge of Chaos NeurIPS 2017 Multi-View Supervision for Single-View Reconstruction via Differentiable Ray Consistency CVPR 2017 Nonlinear random matrix theory for deep learning NeurIPS 2017 On Approximation Guarantees for Greedy Low Rank Optimization ICML 2017 On the Complexity of Learning Neural Networks NeurIPS 2017 On the Flatness of Loss Surface for Two-layered ReLU Networks ACML 2017 On the Learnability of Fully-Connected Neural Networks AISTATS 2017 Principles of Riemannian Geometry in Neural Networks NeurIPS 2017 Resurrecting the sigmoid in deep learning through dynamical isometry: theory and practice NeurIPS 2017 SafetyNets: Verifiable Execution of Deep Neural Networks on an Untrusted Cloud NeurIPS 2017 SGD Learns the Conjugate Kernel Class of the Network NeurIPS 2017