convergence analysis
A study of the conditions under which the output of an algorithm approaches a specific value or solution as the number of iterations increases. In AI, especially in optimization and learning algorithms, it helps determine whether a model will reach an optimal solution.
- A Principled Path to Fitted Distributional Evaluation
- A Single-Loop Gradient Algorithm for Pessimistic Bilevel Optimization via Smooth Approximation
- A Unified Analysis of Stochastic Gradient Descent with Arbitrary Data Permutations and Beyond
- Accelerated Vertical Federated Adversarial Learning through Decoupling Layer-Wise Dependencies
- Adaptive Sigmoid Clipping for Balancing the Direction–Magnitude Mismatch Trade-off in Differentially Private Learning
- Breaking AR’s Sampling Bottleneck: Provable Acceleration via Diffusion Language Models
- Efficient Utility-Preserving Machine Unlearning with Implicit Gradient Surgery
- Efficiently Escaping Saddle Points under Generalized Smoothness via Self-Bounding Regularity
- Fast Last-Iterate Convergence of SGD in the Smooth Interpolation Regime
- Graph-based Symbolic Regression with Invariance and Constraint Encoding
- Joint Hierarchical Representation Learning of Samples and Features via Informed Tree-Wasserstein Distance
- Meta-D2AG: Causal Graph Learning with Interventional Dynamic Data
- Multi-Class Support Vector Machine with Differential Privacy
- Non-Singularity of the Gradient Descent Map for Neural Networks with Piecewise Analytic Activations
- Personalized Subgraph Federated Learning with Differentiable Auxiliary Projections
- Precise Diffusion Inversion: Towards Novel Samples and Few-Step Models
- Private Evolution Converges
- REINFORCE Converges to Optimal Policies with Any Learning Rate
- Robust Distributed Estimation: Extending Gossip Algorithms to Ranking and Trimmed Means
- SPFL: Sequential updates with Parallel aggregation for Enhanced Federated Learning under Category and Domain Shifts
- Sketched Adaptive Distributed Deep Learning: A Sharp Convergence Analysis
- Stable Coresets via Posterior Sampling: Aligning Induced and Full Loss Landscapes
- Theoretical Investigation of Adafactor for Non-Convex Smooth Optimization