The integration of discrete algorithmic components in deep learning architectures has numerous applications. Recently, Implicit Maximum Likelihood Estimation, a class gradient estimators for exponential family distributions, was proposed by combining implicit differentiation through perturbation with the path-wise estimator. However, due to finite difference approximation gradients, it is espec...

It is shown (Proposition (3.9)) that the asymptotic information bound which is valid for the estimation of a parameter in the structure (mixture) model remains valid in the functional model (incidental nuisance parameters) if only perturbation symmetric estimators (Deenition (3.6)) are admitted. Perturbation symmetry is a property which is closely related to permutation symmetry (Theorem (3.4))...

We extend previous works of Fliess et al. [2008] on the estimation of the switching signal and of the state for switching linear systems to the perturbed case when the perturbation is structured that is when the perturbation is unknown but known to satisfy a certain differential equation (for example if the perturbation is constant then its time-derivative is zero). We characterize also singula...

We address the problem of determining whether or not a harmonic oscillator has been perturbed by an external force. Quantum detection and estimation theory has been used in devising optimum measurement schemes. Detection probability has been evaluated for different initial state preparations of oscillator. The corresponding lower bounds on minimum detectable perturbation intensity has been eval...

DTrj QUXIAJAT flSCFTED 3 The problem of linearly constrained least squares has many applications in signal processing. In this paper, we present a perturbation analysis of a linearly constrrined least squares algorithm for adaptive beaniforming. The perturbation bounds for the solution as well as for the alaest residual element are derived. We also propose an error estimation scheme for the res...

This paper studies the Schatten-q error of low-rank matrix estimation by singular value decomposition under perturbation. We specifically establish a perturbation bound on via projection bound. Then, we lower bounds to justify tightness upper error. further develop user-friendly sin? for subspace based Finally, demonstrate advantage our results over ones in literature simulation.

The authors consider the impact of a small but nonzero bandwidth on narrowband direction-of-arrival (DOA) estimation using an array of sensors. They derive expressions for the DOA error for three commonly used array processing algorithms: MUSIC, ESPRIT and weighted subspace fitting (WSF). The error expressions are found by a perturbation analysis of these algorithms for small relative bandwidth...

The paper deals with the local sensitivity analysis of the discrete-time infinitehorizon H∞ estimation problem. Both linear and nonlinear perturbation bounds are derived for the solution of the matrix Riccati equation that determines the sensitivity of the problem.

This paper is concerned with gradient estimation techniques for steady-state performance measures associated with regenerative stochastic processes. The principal emphasis is on the discussion of conditions under which likelihood ratio methods and infinitesimal perturbation analysis techniques are valid.

This letter proposes a multiple parametric dictionary learning algorithm for direction of arrival (DOA) estimation in presence of array gain-phase error and mutual coupling. It jointly solves both the DOA estimation and array imperfection problems to yield a robust DOA estimation in presence of array imperfection errors and off-grid. In the proposed method, a multiple parametric dictionary lear...