Theory

A Depth Hierarchy for Computing the Maximum in ReLU Networks via Extremal Graph Theory COLT 2026 Advances in PAC-Bayesian certification of deep neural networks: tighter closed-form inequalities and optimization of bounds on non-differentiable losses UAI 2026 A Functional-Space Mean-Field Theory of Partially-Trained Three-Layer Neural Networks JMLR 2026 A Mean-Field Analysis of Neural Stochastic Gradient Descent-Ascent for Functional Minimax Optimization JMLR 2026 Attention Sinks Are Provably Necessary in Softmax Transformers: Evidence from Trigger-Conditional Tasks ACL 2026 Beyond Unconstrained Features: Neural Collapse for Shallow Neural Networks with General Data JMLR 2026 C^2FG: Control Classifier-Free Guidance via Score Discrepancy Analysis CVPR 2026 Characterizing the Expressivity of Local Attention in Transformers ACL 2026 Error Analysis for Deep ReLU Feedforward Density-Ratio Estimation with Bregman Divergence JMLR 2026 Flavors of Margin: Implicit Bias of Steepest Descent in Homogeneous Neural Networks JMLR 2026 High-Dimensional Analysis of Gradient Flow for Extensive-Width Quadratic Neural Networks JMLR 2026 How Does the ReLU Activation Affect the Implicit Bias of Gradient Descent on High-dimensional Neural Network Regression? COLT 2026 InTrain: Intrinsic Trainability for Zero-Cost Neural Architecture Search CVPR 2026 Invited Open Problem: Is the Power of Deep Learning over Linear Models Inherently Distribution Dependent? COLT 2026 Is the Attention Matrix Really the Key to Self-Attention in Multivariate Long-Term Time Series Forecasting? ACL 2026 Learning Expressive Random Feature Models via Parametrized Activations UAI 2026 Neural Differentiation in Deep Networks: A Theoretical Framework for Expressivity and Representational Diversity CVPR 2026 Nonlocal Techniques for the Analysis of Deep ReLU Neural Network Approximations JMLR 2026 On the Convergence of Overparameterized Problems: Inherent Properties of the Compositional Structure of Neural Networks L4DC 2026 On the Equivalence of Random Network Distillation, Deep Ensembles, and Bayesian Inference UAI 2026 Optimal Approximation and Generalization Errors for Deep Convolutional Neural Networks JMLR 2026 Optimal Neural Network Approximation of Smooth Compositional Functions on Sets with Low Intrinsic Dimension COLT 2026 POUR: A Provably Optimal Method for Unlearning Representation via Neural Collapse CVPR 2026 Price of metric universality in vector quantization is at most 0.11 bit COLT 2026 Provable Learning of Random Hierarchy Models and Hierarchical Shallow-to-Deep Chaining COLT 2026