In this paper we investigate the problem of predicting the closed set identification performance of biometric matchers in large-scale applications given their corresponding performances in small-scale applications. We identify two major effects responsible for the prediction errors in previously proposed methods: the binomial approximation effect and the score mixing effect. We propose to use a...

We present an approximation algorithm that takes a pool of pre-trained models as input and produces from it a cascaded model with similar accuracy but lower average-case cost. Applied to state-of-the-art ImageNet classification models, this yields up to a 2x reduction in floating point multiplications, and up to a 6x reduction in average-case memory I/O. The auto-generated cascades exhibit intu...

Our goal is to improve the training and prediction time of Nyström method, which is a widely-used technique for generating low-rank kernel matrix approximations. When applying the Nyström approximation for large-scale applications, both training and prediction time is dominated by computing kernel values between a data point and all landmark points. With m landmark points, this computation requ...

Software performance evaluation relies on the ability of simple models to predict the performance of complex systems. Often, however, the models are not capturing potentially relevant effects in system behavior, such as sharing of memory caches or sharing of cores by hardware threads. The goal of this paper is to investigate whether and to what degree a simple linear adjustment of service deman...

This paper is dedicated to Giorgio Picci on the occasion of his 65th birthday. I have come to appreciate Giorgio not only as a great friend but also as a great scholar. When we first met at Brown University in 1973, he introduced me to his seminal paper [29] on splitting subspaces, which became the impetus for our joint work on the geometric theory of linear stochastic systems [23–26]. This led...

in this work we deal with the question: how can one improve the approximation level for some nonlinear integral equations? good candidates for this aim are semi orthogonal b-spline scaling functions and their duals. although there are different works in this area, only b-spline of degree at most 2 are used for this approximation. here we compute b-spline scaling functions of degree 4 and their ...

An analytical solution is presented that reconstructs residual stress field from limited and incomplete data.  The inverse problem of reconstructing residual stresses is solved using an appropriate form of the airy stress function.  This function is chosen to satisfy the stress equilibrium equations together with the boundary conditions for a domain within a convex polygon.  The analytical solu...

Let $L$ be an integral and commutative quantale. In this paper, by fuzzifying the notion of generalized neighborhood systems, the notion of $L$-fuzzy generalized neighborhoodsystem is introduced and then a pair of lower and upperapproximation operators based on it are defined and discussed. It is proved that these approximation operators include generalized neighborhood system...