ICML 2020 accepted papers on Learning theory and algorithms
220 of the 1,087 papers accepted at ICML 2020 (20.2%) are on Learning theory and algorithms, compared with 18.9% at ICML 2019. They fall into 9 topics.
See ICML 2020 in the interactive map
Submitting to ICML? 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 · 52 papers
- Problems with Shapley-value-based explanations as feature importance measures (110 citations)
- Searching to Exploit Memorization Effect in Learning with Noisy Labels (53 citations)
- DeltaGrad: Rapid retraining of machine learning models (42 citations)
- Decision Trees for Decision-Making under the Predict-then-Optimize Framework (41 citations)
- Generalized and Scalable Optimal Sparse Decision Trees (41 citations)
- Feature Selection using Stochastic Gates (40 citations)
- The Many Shapley Values for Model Explanation (36 citations)
- Fairwashing explanations with off-manifold detergent (35 citations)
- Learning Reasoning Strategies in End-to-End Differentiable Proving (30 citations)
- Predictive Multiplicity in Classification (29 citations)
- A Distributional Framework For Data Valuation (25 citations)
- Consistent Estimators for Learning to Defer to an Expert (24 citations)
and 40 more in the interactive map.
Sample complexity · 45 papers
- CLUB: A Contrastive Log-ratio Upper Bound of Mutual Information (57 citations)
- Curse of Dimensionality on Randomized Smoothing for Certifiable Robustness (33 citations)
- Linear Lower Bounds and Conditioning of Differentiable Games (25 citations)
- Robust One-Bit Recovery via ReLU Generative Networks: Near-Optimal Statistical Rate and Global Landscape Analysis (14 citations)
- Optimal Bounds between f-Divergences and Integral Probability Metrics (13 citations)
- Quantum Expectation-Maximization for Gaussian mixture models (13 citations)
- Logistic Regression for Massive Data with Rare Events (11 citations)
- Acceleration through spectral density estimation (8 citations)
- On Relativistic f-Divergences (6 citations)
- On the Power of Compressed Sensing with Generative Models (6 citations)
- Provable guarantees for decision tree induction: the agnostic setting (6 citations)
- Data Amplification: Instance-Optimal Property Estimation (5 citations)
and 33 more in the interactive map.
Kernels and regression theory · 30 papers
- An Investigation of Why Overparameterization Exacerbates Spurious Correlations (106 citations)
- Generalisation error in learning with random features and the hidden manifold model (100 citations)
- Rethinking Bias-Variance Trade-off for Generalization of Neural Networks (77 citations)
- Learning Deep Kernels for Non-Parametric Two-Sample Tests (75 citations)
- Double Trouble in Double Descent: Bias and Variance(s) in the Lazy Regime (48 citations)
- The Neural Tangent Kernel in High Dimensions: Triple Descent and a Multi-Scale Theory of Generalization (32 citations)
- Spectrum Dependent Learning Curves in Kernel Regression and Wide Neural Networks (27 citations)
- Normalized Flat Minima: Exploring Scale Invariant Definition of Flat Minima for Neural Networks Using PAC-Bayesian Analysis (13 citations)
- Small Data, Big Decisions: Model Selection in the Small-Data Regime (10 citations)
- Implicit Regularization of Random Feature Models (8 citations)
- Near-Tight Margin-Based Generalization Bounds for Support Vector Machines (8 citations)
- Near Input Sparsity Time Kernel Embeddings via Adaptive Sampling (6 citations)
and 18 more in the interactive map.
Matrix and tensor methods · 27 papers
- Random Matrix Theory Proves that Deep Learning Representations of GAN-data Behave as Gaussian Mixtures (18 citations)
- High-dimensional Robust Mean Estimation via Gradient Descent (14 citations)
- Message Passing Least Squares Framework and its Application to Rotation Synchronization (12 citations)
- Learning Algebraic Multigrid Using Graph Neural Networks (10 citations)
- Uncertainty quantification for nonconvex tensor completion: Confidence intervals, heteroscedasticity and optimality (9 citations)
- Adaptive Sketching for Fast and Convergent Canonical Polyadic Decomposition (6 citations)
- Learning Factorized Weight Matrix for Joint Filtering (6 citations)
- Schatten Norms in Matrix Streams: Hello Sparsity, Goodbye Dimension (5 citations)
- When deep denoising meets iterative phase retrieval (5 citations)
- Polynomial Tensor Sketch for Element-wise Function of Low-Rank Matrix (4 citations)
- Random extrapolation for primal-dual coordinate descent (4 citations)
- Tensor denoising and completion based on ordinal observations (4 citations)
and 15 more in the interactive map.
Combinatorial and search optimization · 16 papers
- Lifted Disjoint Paths with Application in Multiple Object Tracking (89 citations)
- Retro*: Learning Retrosynthetic Planning with Neural Guided A* Search (42 citations)
- Efficient Continuous Pareto Exploration in Multi-Task Learning (31 citations)
- Optimizer Benchmarking Needs to Account for Hyperparameter Tuning (26 citations)
- On the Global Optimality of Model-Agnostic Meta-Learning (15 citations)
- Learning to Simulate and Design for Structural Engineering (12 citations)
- Safe screening rules for L0-regression from Perspective Relaxations (12 citations)
- ECLIPSE: An Extreme-Scale Linear Program Solver for Web-Applications (10 citations)
- It’s Not What Machines Can Learn, It’s What We Cannot Teach (8 citations)
- Neural Networks are Convex Regularizers: Exact Polynomial-time Convex Optimization Formulations for Two-layer Networks (7 citations)
- Learning for Dose Allocation in Adaptive Clinical Trials with Safety Constraints (6 citations)
- Learning Optimal Tree Models under Beam Search (6 citations)
and 4 more in the interactive map.
Fairness · 15 papers
- Minimax Pareto Fairness: A Multi Objective Perspective (54 citations)
- Is There a Trade-Off Between Fairness and Accuracy? A Perspective Using Mismatched Hypothesis Testing (48 citations)
- Data preprocessing to mitigate bias: A maximum entropy based approach (36 citations)
- Collaborative Machine Learning with Incentive-Aware Model Rewards (31 citations)
- Measuring Non-Expert Comprehension of Machine Learning Fairness Metrics (28 citations)
- Fair Learning with Private Demographic Data (26 citations)
- Two Simple Ways to Learn Individual Fairness Metrics from Data (24 citations)
- DeBayes: a Bayesian Method for Debiasing Network Embeddings (21 citations)
- FACT: A Diagnostic for Group Fairness Trade-offs (21 citations)
- FR-Train: A Mutual Information-Based Approach to Fair and Robust Training (13 citations)
- Fair k-Centers via Maximum Matching (11 citations)
- A Pairwise Fair and Community-preserving Approach to k-Center Clustering (10 citations)
and 3 more in the interactive map.
Submodular and game algorithms · 15 papers
- Do GANs always have Nash equilibria? (30 citations)
- Invariant Risk Minimization Games (29 citations)
- Streaming Submodular Maximization under a k-Set System Constraint (16 citations)
- Optimizing Long-term Social Welfare in Recommender Systems: A Constrained Matching Approach (11 citations)
- The FAST Algorithm for Submodular Maximization (10 citations)
- Streaming k-Submodular Maximization under Noise subject to Size Constraint (9 citations)
- On the Unreasonable Effectiveness of the Greedy Algorithm: Greedy Adapts to Sharpness (6 citations)
- Optimal approximation for unconstrained non-submodular minimization (6 citations)
- Parallel Algorithm for Non-Monotone DR-Submodular Maximization (6 citations)
- Converging to Team-Maxmin Equilibria in Zero-Sum Multiplayer Games (5 citations)
- Fast and Private Submodular and $k$-Submodular Functions Maximization with Matroid Constraints (5 citations)
- Sparsified Linear Programming for Zero-Sum Equilibrium Finding (5 citations)
and 3 more in the interactive map.
Clustering · 13 papers
- Spectral Clustering with Graph Neural Networks for Graph Pooling (76 citations)
- Individual Fairness for k-Clustering (31 citations)
- Explainable k-Means and k-Medians Clustering (20 citations)
- Sparse Subspace Clustering with Entropy-Norm (14 citations)
- Coresets for Clustering in Graphs of Bounded Treewidth (7 citations)
- Expert Learning through Generalized Inverse Multiobjective Optimization: Models, Insights, and Algorithms (7 citations)
- Deep Divergence Learning (5 citations)
- k-means++: few more steps yield constant approximation (5 citations)
- Simple and sharp analysis of k-means|| (5 citations)
- On hyperparameter tuning in general clustering problemsm (4 citations)
- Sets Clustering (3 citations)
- Correlation Clustering with Asymmetric Classification Errors (2 citations)
and 1 more in the interactive map.
Optimal transport · 7 papers
- Missing Data Imputation using Optimal Transport (43 citations)
- On Unbalanced Optimal Transport: An Analysis of Sinkhorn Algorithm (24 citations)
- Adversarial Risk via Optimal Transport and Optimal Couplings (23 citations)
- Debiased Sinkhorn barycenters (18 citations)
- A Swiss Army Knife for Minimax Optimal Transport (10 citations)
- Scalable Nearest Neighbor Search for Optimal Transport (10 citations)
- Regularized Optimal Transport is Ground Cost Adversarial (9 citations)
