Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 17 of 17 for “"exact gradient"”.
-
Recovery Techniques For Finite Element Methods And Their Applications
… but also to provide an asymptotically exact posteriori error estimators for adaptive meth- ods. This dissertation presents recovery techniques for nonconforming finite element methods and high order derivative as well as applications of gradient recovery. Our first target is to develop …
-
A decentralised graph-based framework for electrical power markets
… links in the solution process while allowing the exact gradient to be computed. Following, recognizing that the DC optimal power flow is a quadratic separable program, this model is generalised to a quadratic separable program model. Finally, an agent oriented approach is proposed in order to …
-
Design, Analysis, and Fabrication of a Snake-Inspired Robot with a Rectilinear Gait
… formulation for joint trajectory generation. Exact gradient computation of the torque functions is presented. A parametric model is used to describe the performance effects of changing system parameters such as mass, length, and motor speed. Finally, a snake-inspired robot is designed and …
-
Stochastic Gradient Methods with Bias and Momentum
… stochastic optimization methods where exact gradient computations are prohibitively expensive, necessitating potentially biased gradient estimators. Such challenges arise in applications like empirical risk minimization in machine learning, where unbiased gradient estimates are not …
-
Hamiltonian Monte Carlo for Reconstructing Historical Earthquake-Induced Tsunamis
… reduce computation time. Unfortunately the exact gradient for this problem is not available, and so we make use of a surrogate gradient via a neural network fitted to the forward model. We examine the effects of this surrogate gradient HMC sampling method on the posterior distribution for an …
-
Reconstructing Historical Earthquake-Induced Tsunamis: Case Study of 1852 Event Using the Adjoint Method Combined with HMC
… such as Hamiltonian Monte Carlo (HMC) where the gradient of the forward model (Geoclaw) is required. This is problematic however as this gradient is not available computationally. To mitigate this problem, we make use of a linearized adjoint solver for the shallow water equations, and exact …
-
Statistical learning of Hawkes models and market microstructure
… error (LSE), as well as the evaluation of their gradients, have a quadratic complexity in the number of observed events. In practice, this prohibits the use of exact gradient-based algorithms for parameter estimation in many settings. Furthermore, MHP models are not designed for non-stationary …
-
Numerical algorithms for differential equations with periodicity
… polynomials and optimization techniques with exact gradient and exact Hessian matrix. New choreographies in spaces of constant curvature are found.</p> <p>Exponential integrators for solving periodic semilinear stiff PDEs in 1D, 2D and 3D periodic domains are reviewed, and 30 exponential …
-
Antenna Array Output Power Minimization Using Steepest Descent Adaptive Algorithm
… algorithm, named the Hilbert-space-based (HSB) gradient method, is based on the steepest descent algorithm and implements an efficient, exact gradient calculation. With M antennas in the array, only M − 1 weights are adjustable; one antenna weight is held constant to prevent the algorithm from …
-
Συζυγείς μέθοδοι για το σχεδιασμό μορφών με βέλτιστη αεροδυναμική συμπεριφορά σε στρωτές και τυρβώδεις ροές
… The steepest descent method, supported by the exact gradient of the cost function, is used whenever an optimization problem is solved. Case studies related to turbomachinery blade airfoils and duct flows are presented. In addition, in the context of collaboration of the Parallel CFD & …
-
Fast methods for fitting log-Gaussian Cox process models in ecology.
… differentiation that enables us to obtain exact gradient information rapidly for computationally efficient optimisation and inference. We demonstrate the method’s performance through both simulations and a real data application, with promising results in terms of computational speed and …
-
High-Order Finite-Volume Schemes for Magnetohydrodynamics
… approach is employed, which combines unlimited k-exact polynomial reconstruction with a monotonicity preserving scheme. The CENO schemes, which were originally developed for compressible fluid flow, are applied to the MHD equations, along with two possible control mechanisms for divergence error …
-
Proxy Gradient Estimation for Black-Box Feature Attribution
… that can exploit full model details, such as exact gradient measurements. To bridge the performance gap between black-box and white-box feature attribution methods, this dissertation introduces a novel framework based on gradient estimation, which is then extended to a more general formulation …
-
Particle methods for Bayesian inverse problems governed by partial differential equations (PDEs)
… of forward solves or require the evaluation of gradients or Hessians, which may not be available. Even when gradient and Hessian information is available, computing this derivative information is computationally expensive. We present the Sequential Ensemble Transport (SET) and the Optimization …
-
Simulation and analysis of gradient coding protocols- Leveraging partial stragglers within gradient coding
… the parameter server to wait indefinitely for gradient computations, severely degrading training efficiency. Gradient coding, a coding-theoretic approach to distributed learning, addresses this problem by introducing redundancy in the assignment of data chunks to workers. By carefully encoding …
-
Robust model predictive control of robots in confined spaces
… constraint formulation allows straightforward exact gradient calculation and real-time feasibility of the resulting controller.<br><br>The second part of the thesis ensures robust collision avoidance in the presence of disturbances from model-plant mismatch and uncertainties in the movement of …
-
Stochastic algorithms for distributed optimization and machine learning
… devoted to investigating randomized incremental gradient (RIG) methods for solving the class of finite-sum optimization problems, which is the core problem in distributed optimization. By developing and proving the optimality of the primal-dual gradient method, which covers the well-known …