ICLR 2019 accepted papers on Learning theory and algorithms
37 of the 502 papers accepted at ICLR 2019 (7.4%) are on Learning theory and algorithms, compared with 7.1% at ICLR 2018. They fall into 6 topics.
See ICLR 2019 in the interactive map
Submitting to ICLR? 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
Prediction and decision losses · 12 papers
- Efficient Lifelong Learning with A-GEM (641 citations)
- Learning-Based Frequency Estimation Algorithms (65 citations)
- Neural Persistence: A Complexity Measure for Deep Neural Networks Using Algebraic Topology (61 citations)
- Janossy Pooling: Learning Deep Permutation-Invariant Functions for Variable-Size Inputs (58 citations)
- On the loss landscape of a class of deep neural networks with no bad local valleys (56 citations)
- L-Shapley and C-Shapley: Efficient Model Interpretation for Structured Data (37 citations)
- Small nonlinearities in activation functions create bad local minima in neural networks (26 citations)
- Towards Robust, Locally Linear Deep Networks (25 citations)
- Attentive Neural Processes (24 citations)
- Conditional Network Embeddings (23 citations)
- Caveats for information bottleneck in deterministic scenarios (18 citations)
- Unsupervised Learning of the Set of Local Maxima (6 citations)
Matrix and tensor methods · 8 papers
- ALISTA: Analytic Weights Are As Good As Learned Weights in LISTA (98 citations)
- A Data-Driven and Distributed Approach to Sparse Signal Representation and Recovery (20 citations)
- Learning Grid Cells as Vector Representation of Self-Position Coupled with Matrix Representation of Self-Motion (12 citations)
- NOODL: Provable Online Dictionary Learning and Sparse Coding (10 citations)
- Random mesh projectors for inverse problems (8 citations)
- A Kernel Random Matrix-Based Approach for Sparse PCA (7 citations)
- Efficient Training on Very Large Corpora via Gramian Estimation (7 citations)
- Learning Two-layer Neural Networks with Symmetric Inputs (4 citations)
Kernels and regression theory · 7 papers
- Meta-learning with differentiable closed-form solvers (149 citations)
- The role of over-parametrization in generalization of neural networks (144 citations)
- Deep learning generalizes because the parameter-function map is biased towards simple functions (110 citations)
- Adaptivity of deep ReLU network for learning in Besov and mixed smooth Besov spaces: optimal rate and curse of dimensionality (105 citations)
- Non-vacuous Generalization Bounds at the ImageNet Scale: a PAC-Bayesian Compression Approach (88 citations)
- An analytic theory of generalization dynamics and transfer learning in deep linear networks (50 citations)
- Data-Dependent Coresets for Compressing Neural Networks with Applications to Generalization Bounds (5 citations)
Combinatorial and search optimization · 4 papers
- Learning To Solve Circuit-SAT: An Unsupervised Differentiable Approach (45 citations)
- Neural Logic Machines (45 citations)
- Detecting Egregious Responses in Neural Sequence-to-sequence Models (25 citations)
- Optimal Completion Distillation for Sequence Learning (5 citations)
Sample complexity · 4 papers
- Approximability of Discriminators Implies Diversity in GANs (35 citations)
- Deep, Skinny Neural Networks are not Universal Approximators (6 citations)
- ROBUST ESTIMATION VIA GENERATIVE ADVERSARIAL NETWORKS (3 citations)
- The Comparative Power of ReLU Networks and Polynomial Kernels in the Presence of Sparse Latent Structure (2 citations)
Optimal transport · 2 papers
- Scalable Unbalanced Optimal Transport using Generative Adversarial Networks (13 citations)
- Wasserstein Barycenter Model Ensembling (13 citations)
