neural network training
The process of adjusting the parameters of a neural network model by minimizing a loss function using optimization techniques, typically involving backpropagation and gradient descent.
- Error Broadcast and Decorrelation as a Potential Artificial and Natural Learning Mechanism
- Escaping saddle points without Lipschitz smoothness: the power of nonlinear preconditioning
- Improving the Straight-Through Estimator with Zeroth-Order Information
- KOALA++: Efficient Kalman-Based Optimization with Gradient-Covariance Products
- Learn2Mix: Training Neural Networks Using Adaptive Data Integration
- Mean Flows for One-step Generative Modeling
- Mean Flows for One-step Generative Modeling
- Nonlinearly Preconditioned Gradient Methods: Momentum and Stochastic Analysis
- On the Surprising Effectiveness of Large Learning Rates under Standard Width Scaling
- Uncertainty-Guided Exploration for Efficient AlphaZero Training