NeurIPS 2017 accepted papers on Optimization
71 of the 688 papers accepted at NeurIPS 2017 (10.3%) are on Optimization, compared with 11.1% at NeurIPS 2016. They fall into 4 topics.
See NeurIPS 2017 in the interactive map
Submitting to NeurIPS? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Nonconvex and smooth optimization · 35 papers
- Non-convex Finite-Sum Optimization Via SCSG Methods (129 citations)
- Variance-based Regularization with Convex Objectives (104 citations)
- Gradient Descent Can Take Exponential Time to Escape Saddle Points (63 citations)
- Asynchronous Coordinate Descent under More Realistic Assumptions (42 citations)
- Convergent Block Coordinate Descent for Training Tikhonov Regularized Deep Neural Networks (41 citations)
- Linear Convergence of a Frank-Wolfe Type Algorithm over Trace-Norm Balls (34 citations)
- Gradient Methods for Submodular Maximization (27 citations)
- Integration Methods and Optimization Algorithms (26 citations)
- Dykstra's Algorithm, ADMM, and Coordinate Descent: Connections, Insights, and Extensions (23 citations)
- On Frank-Wolfe and Equilibrium Computation (23 citations)
- Differentially Private Empirical Risk Minimization Revisited: Faster and More General (21 citations)
- Doubly Accelerated Stochastic Variance Reduced Dual Averaging Method for Regularized Empirical Risk Minimization (20 citations)
and 23 more in the interactive map.
Optimizers · 21 papers
- GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium (4,380 citations)
- The Marginal Value of Adaptive Gradient Methods in Machine Learning (546 citations)
- The Numerics of GANs (175 citations)
- Stein Variational Gradient Descent as Gradient Flow (112 citations)
- Gradient descent GAN optimization is locally stable (71 citations)
- Multiscale Quantization for Fast Similarity Search (46 citations)
- Reducing Reparameterization Gradient Variance (34 citations)
- Diving into the shallows: a computational perspective on large-scale shallow learning (31 citations)
- VAE Learning via Stein Variational Gradient Descent (18 citations)
- Scalable Planning with Tensorflow for Hybrid Nonlinear Domains (12 citations)
- Stochastic Submodular Maximization: The Case of Coverage Functions (12 citations)
- Fast Black-box Variational Inference through Stochastic Trust-Region Optimization (7 citations)
and 9 more in the interactive map.
Implicit bias and generalization · 8 papers
- Train longer, generalize better: closing the generalization gap in large batch training of neural networks (399 citations)
- Convergence Analysis of Two-layer Neural Networks with ReLU Activation (270 citations)
- Implicit Regularization in Matrix Factorization (135 citations)
- Regularizing Deep Neural Networks by Noise: Its Interpretation and Optimization (134 citations)
- Learning ReLUs via Gradient Descent (88 citations)
- SGD Learns the Conjugate Kernel Class of the Network (69 citations)
- On Structured Prediction Theory with Calibrated Convex Surrogate Losses (9 citations)
- Convolutional Phase Retrieval
Distributed and compressed · 7 papers
- Machine Learning with Adversaries: Byzantine Tolerant Gradient Descent (967 citations)
- QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding (911 citations)
- TernGrad: Ternary Gradients to Reduce Communication in Distributed Deep Learning (662 citations)
- Can Decentralized Algorithms Outperform Centralized Algorithms? A Case Study for Decentralized Parallel Stochastic Gradient Descent (422 citations)
- Collaborative Deep Learning in Fixed Topology Networks (115 citations)
- Straggler Mitigation in Distributed Optimization Through Data Encoding (109 citations)
- Communication-Efficient Stochastic Gradient Descent, with Applications to Neural Networks (4 citations)
