A procedure to approximate non-causal adaptive IIR filtering is described. Such structures could be useful, for instance, in the equalization of non-minimum phase communication channels. The procedure is based on backward calculations within blocks of samples and we show that for adaptive filtering the overlap-save method is more suitable than the overlap-add method. Moreover, a temporary freez...

Many large MDPs can be represented compactly using a dynamic Bayesian network. Although the structure of the value function does not re­ tain the structure of the process, recent work has suggested that value functions in factored MDPs can often be approximated well using a factored value function: a linear combination of restr icted basis functions, each of which refers only to a small subset ...

We derive computable expressions of structured backward errors of approximate eigenelements of ∗-palindromic and ∗-anti-palindromic matrix polynomials. We also characterize minimal structured perturbations such that approximate eigenelements are exact eigenelements of the perturbed polynomials. We detect structure preserving linearizations which have almost no adverse effect on the structured b...

We present two “fast” approaches to the NP-hard problem of computing a maximally sparse approximate solution to linear inverse problems, also known as best subset selection. The first approach, a heuristic, is an iterative algorithm globally convergent to sparse elements of any given convex, compact S C Wmr. We demonstrate its effectiveness in bandlimited extrapolation and in sparse filter desi...

Second-order optimization methods have the ability to accelerate convergence by modifying gradient through curvature matrix. There been many attempts use second-order for training deep neural networks. In this work, inspired diagonal approximations and factored such as Kronecker-factored Approximate Curvature (KFAC), we propose a new approximation Fisher information matrix (FIM) called Trace-re...

This paper presents and analyzes a new algorithm for computing eigenvectors of symmetric tridiagonal matrices factored as LDLt, with D diagonal and L unit bidiagonal. If an eigenpair is well behaved in a certain sense with respect to the factorization, the algorithm is shown to compute an approximate eigenvector which is accurate to working precision. As a consequence, all the eigenvectors comp...

Many large MDPs can be represented compactly using a dynamic Bayesian network. Although the structure of the value function does not retain the structure of the process, recent work has suggested that value functions in factored MDPs can often be approximated well using a factored value function: a linear combination of restricted basis functions, each of which refers only to a small subset of ...

We propose a unifying algorithm for non-smooth non-convex optimization. The algorithm approximates the objective function by a convex model function and finds an approximate (Bregman) proximal point of the convex model. This approximate minimizer of the model function yields a descent direction, along which the next iterate is found. Complemented with an Armijo-like line search strategy, we obt...