non-convex objectives
Non-convex objectives refer to optimization problems where the loss function has multiple local minima, making it more challenging to find the global optimum. This is common in deep learning applications.
- Efficient Adaptive Federated Optimization
- Fast Projection-Free Approach (without Optimization Oracle) for Optimization over Compact Convex Set
- On the Optimal Construction of Unbiased Gradient Estimators for Zeroth-Order Optimization
- Solving Neural Min-Max Games: The Role of Architecture, Initialization & Dynamics
- Tree-Guided Diffusion Planner