Finite Difference Gradient Approximation: To Randomize or Not?
From MaRDI portal
Recommendations
- Minimax efficient finite-difference stochastic gradient estimators using black-box function evaluations
- Stochastic adversarial noise in the ``black box optimization problem
- On the numerical performance of finite-difference-based methods for derivative-free optimization
- A mixed finite differences scheme for gradient approximation
- Random gradient-free minimization of convex functions
Cites work
- A derivative-free trust-region algorithm for the optimization of functions smoothed via Gaussian convolution using adaptive multiple importance sampling
- A theoretical and empirical comparison of gradient approximations in derivative-free optimization
- Adaptive stochastic approximation by the simultaneous perturbation method
- An Optimal Algorithm for Bandit and Zero-Order Convex Optimization with Two-Point Feedback
- Derivative-free optimization methods
- Estimating derivatives of noisy simulations
- Multivariate stochastic approximation using a simultaneous perturbation gradient approximation
- Online convex optimization in the bandit setting: gradient descent without a gradient
- Optimal Rates for Zero-Order Convex Optimization: The Power of Two Function Evaluations
- Random gradient-free minimization of convex functions
- Spearmint
- Stochastic First- and Zeroth-Order Methods for Nonconvex Stochastic Programming
Cited in
(10)- Minimax efficient finite-difference stochastic gradient estimators using black-box function evaluations
- Gradient-free federated learning methods with l₁ and l₂-randomization for non-smooth convex stochastic optimization problems
- Small errors in random zeroth-order optimization are imaginary
- Zeroth-order Riemannian averaging stochastic approximation algorithms
- The ``black-box optimization problem: zero-order accelerated stochastic method via kernel approximation
- Accelerated gradient methods with biased gradient estimates: risk sensitivity, high-probability guarantees, and large deviation bounds
- Derivative-free optimization with transformed objective functions and the algorithm based on the least Frobenius norm updating quadratic model
- Subdifferentially polynomially bounded functions and Gaussian smoothing-based zeroth-order optimization
- Stochastic ISTA/FISTA adaptive step search algorithms for convex composite optimization
- Globally convergent derivative-free methods in nonconvex optimization with and without noise
This page was built for publication: Finite Difference Gradient Approximation: To Randomize or Not?
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5057983)
