generalization guarantees
The assurances that a trained model will perform well on unseen data, crucial for validating its robustness and usability in real-world applications.
- Adaptive Data Analysis for Growing Data
- C3PO: Optimized Large Language Model Cascades with Probabilistic Cost Constraints for Reasoning
- FedFACT: A Provable Framework for Controllable Group-Fairness Calibration in Federated Learning
- Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel
- Kernel Regression in Structured Non-IID Settings: Theory and Implications for Denoising Score Learning
- Marginal-Nonuniform PAC Learnability
- Multi-head Transformers Provably Learn Symbolic Multi-step Reasoning via Gradient Descent
- Rewind-to-Delete: Certified Machine Unlearning for Nonconvex Functions
- Tighter CMI-Based Generalization Bounds via Stochastic Projection and Quantization
- Tighter CMI-Based Generalization Bounds via Stochastic Projection and Quantization