On decompositional algorithms for uniform sampling from n-spheres and n-balls

Critical behavior of the three-dimensional random-anisotropy Heisenberg model

We have studied the critical properties of the three-dimensional random anisotropy Heisenberg model by means of numerical simulations using the Parallel Tempering method. We have simulated the model with two different disorder distributions, cubic and isotropic ones, with two different anisotropy strengths for each disorder class. For the case of the anisotropic disorder, we have found evidence of universality by finding critical exponents and universal dimensionless ratios independent of the strength of the disorder. In the case of isotropic disorder distribution the situation is very involved: we have found two phase transitions in the magnetization channel which are merging for larger lattices remaining a zero magnetization low-temperature phase. Studying this region using a spin-glass order parameter we have found evidence for a spin-glass phase transition. We have estimated effective critical exponents for the spin-glass phase transition for the different values of the strength of the isotropic disorder, discussing the crossover regime.

Statistical modeling of ultrasound signals related to the packing factor of wave scattering phenomena for structural characterization

The two-dimensional homodyned K-distribution has been widely used to model the echo envelope of ultrasound radio frequency (RF) signals in the field of medical ultrasonics. The main contribution of this work is to present a theoretical framework for supporting this model of the echo envelope and statistical models of the RF signals and their Hilbert transform in the case in which the scatterers' positions may be dependent. In doing so, the law of large numbers, Lyapounov's central limit theorem, and the Berry-Esseen theorem are being used. In particular, the proposed theoretical framework supports a previous conjecture relating the scatterer clustering parameter of the homodyned K-distribution to the packing factor W, which is related to the spatial organization of the scatterers, appearing in statistical physics or backscatter coefficient modeling. Simulations showed that the proposed modeling is valid for a number of scatterers and packing factors varying by steps of 2 from 1 to 21 and 1 to 11, respectively. The proposed framework allows, in principle, the detection of the structural information taking place at a scale smaller than the wavelength based solely on the statistical analysis of the RF signals or their echo envelope, although this goal was previously achieved based on the spectral analysis of ultrasound signals.

A robust algorithm for explaining unreliable machine learning survival models using the Kolmogorov-Smirnov bounds

A new robust algorithm based on the explanation method SurvLIME called SurvLIME-KS is proposed for explaining machine learning survival models. The algorithm is developed to ensure robustness to cases of a small amount of training data or outliers of survival data. The first idea behind SurvLIME-KS is to apply the Cox proportional hazards model to approximate the black-box survival model at the local area around a test example due to the linear relationship of covariates in the model. The second idea is to incorporate the well-known Kolmogorov-Smirnov bounds for constructing sets of predicted cumulative hazard functions. As a result, the robust maximin strategy is used, which aims to minimize the average distance between cumulative hazard functions of the explained black-box model and of the approximating Cox model, and to maximize the distance over all cumulative hazard functions in the interval produced by the Kolmogorov-Smirnov bounds. The maximin optimization problem is reduced to the quadratic program. Various numerical experiments with synthetic and real datasets demonstrate the SurvLIME-KS efficiency.

Shape and Spatially-Varying Reflectance Estimation from Virtual Exemplars

This paper addresses the problem of estimating the shape of objects that exhibit spatially-varying reflectance. We assume that multiple images of the object are obtained under a fixed view-point and varying illumination, i.e., the setting of photometric stereo. At the core of our techniques is the assumption that the BRDF at each pixel lies in the non-negative span of a known BRDF dictionary. This assumption enables a per-pixel surface normal and BRDF estimation framework that is computationally tractable and requires no initialization in spite of the underlying problem being non-convex. Our estimation framework first solves for the surface normal at each pixel using a variant of example-based photometric stereo. We design an efficient multi-scale search strategy for estimating the surface normal and subsequently, refine this estimate using a gradient descent procedure. Given the surface normal estimate, we solve for the spatially-varying BRDF by constraining the BRDF at each pixel to be in the span of the BRDF dictionary; here, we use additional priors to further regularize the solution. A hallmark of our approach is that it does not require iterative optimization techniques nor the need for careful initialization, both of which are endemic to most state-of-the-art techniques. We showcase the performance of our technique on a wide range of simulated and real scenes where we outperform competing methods.