generalization bounds
Generalization bounds are theoretical limits that quantify how well a learning algorithm performs on unseen data, providing guarantees that the model trained on a finite dataset will perform adequately on new, real-world data.
- Block Coordinate Descent for Neural Networks Provably Finds Global Minima
- Fast Rate Bounds for Multi-Task and Meta-Learning with Different Sample Sizes
- Generalization Bounds for Kolmogorov-Arnold Networks (KANs) and Enhanced KANs with Lower Lipschitz Complexity
- Generalization Bounds for Rank-sparse Neural Networks
- How Many Domains Suffice for Domain Generalization? A Tight Characterization via the Domain Shattering Dimension
- Kernel Learning with Adversarial Features: Numerical Efficiency and Adaptive Regularization
- Learning to Generalize: An Information Perspective on Neural Processes
- Length Generalization via Auxiliary Tasks
- On the Stability and Generalization of Meta-Learning: the Impact of Inner-Levels
- Synergy over Discrepancy: A Partition-Based Approach to Multi-Domain LLM Fine-Tuning
- Tight High-Probability Bounds for Nonconvex Heavy-Tailed Scenario under Weaker Assumptions
- Which Algorithms Have Tight Generalization Bounds?