Contingent derivatives and regularization for noncoercive inverse problems

Abstract

We study the inverse problem of parameter identification in noncoercive variational problems that commonly appear in applied models. We examine the differentiability of the set-valued parameter-to-solution map using the first-order and the second-order contingent derivatives. We explore the inverse problem using the output least-squares and the modified output least-squares objectives. By regularizing the noncoercive variational problem, we obtain a single-valued regularized parameter-to-solution map and investigate its smoothness and boundedness. We also consider optimization problems using the output least-squares and the modified output least-squares objectives for the regularized variational problem. We give a complete convergence analysis showing that for the output least-squares and the modified output least-squares, the regularized minimization problems approximate the original optimization problems suitably. We also provide the first-order and the second-order adjoint method for the computation of the first-order and the second-order derivatives of the output least-squares objective. We provide discrete formulas for the gradient and the Hessian calculation and present numerical results.     

The survival probability of critical and subcritical branching processes in finite state space Markovian environment

Let ${\left(Z_n\right)}_{n\ge 0}$ be a branching process in a random environment defined by a Markov chain ${\left(X_n\right)}_{n\ge 0}$ with values in a finite state space $\mathbb{X}$. Let ${\mathbb{P}}_i$ be the probability law generated by the trajectories of ${\left(X_n\right)}_{n\ge 0}$ starting at $X_0=i\in \mathbb{X}.$ We study the asymptotic behaviour of the joint survival probability ${\mathbb{P}}_i\left(Z_n>0,X_n=j\right)$, $j\in \mathbb{X}$ as $n\to +\infty$ in the critical and strongly, intermediate and weakly subcritical cases.     

The Bertini irreducibility theorem for higher codimensional slices

In [3], Poonen and Slavov recently developed a novel approach to Bertini irreducibility theorems over an arbitrary field, based on random hyperplane slicing. In this paper, we extend their work by proving an analogous bound for the dimension of the exceptional locus in the setting of linear subspaces of higher codimensions.     

Random walks in cones: The case of nonzero drift

We consider multidimensional discrete valued random walks with nonzero drift killed when leaving general cones of the euclidean space. We find the asymptotics for the exit time from the cone and study weak convergence of the process conditioned on not leaving the cone. We get quasistationarity of its limiting distribution. Finally we construct a version of the random walk conditioned to never leave the cone.     

Multidimensional Scaling With Very Large Datasets

Multidimensional scaling has a wide range of applications when observations are not continuous but it is possible to define a distance (or dissimilarity) among them. However, standard implementations are limited when analyzing very large datasets because they rely on eigendecomposition of the full distance matrix and require very long computing times and large quantities of memory. Here, a new approach is developed based on projection of the observations in a space defined by a subset of the full dataset. The method is easily implemented. A simulation study showed that its performance are satisfactory in different situations and can be run in a short time when the standard method takes a very long time or cannot be run because of memory requirements.     

CARTscans: A Tool for Visualizing Complex Models

We present CARTscans, a graphical tool that displays predicted values across a fourdimensional subspace. We show how these plots are useful for understanding the structure and relationships between variables in a wide variety of models, including (but not limited to) regression trees, ensembles of trees, and linear regressions with varying degrees of interactions. In addition, the common visualization framework allows diverse complex models to be visually compared in a way that illuminates the similarities and differences in the underlying methods, facilitates the choice of a particular model structure, and provides a useful check for implausible predictions of future observations in regions with little or no data.     

Random Variable Generation Using Concavity Properties of Transformed Densities

Abstract

Algorithms are developed for constructing random variable generators for families of densities. The generators depend on the concavity structure of a transformation of the density. The resulting algorithms are rejection algorithms and the methods of this article are concerned with constructing good rejection algorithms for general densities.     

Crystal structure and DFT calculations of methanol {E-N′-(2-hydroxybenzlidene)benzohydrazido}dioxidomolybdenum (VI) complex

The crystal and molecular structure of methanol {E-N′-(2-hydroxybenzlidene)benzohydrazido}dioxidomolybdenum (VI) was determined by single crystal X-ray diffraction. B3LYP/DZP basis set theoretical calculations nicely reproduce the X-ray experimental geometry. The obtained results are discussed in connection with the electronic and structural properties of the compound.     

Predicting USCS soil classification from soil property variables using Random Forest

Soil classification systems are widely used for quickly and easily summarizing soil properties and provide a shorthand method of communication between scientists, engineers, and end-users. Two of the most widely used soil classification systems are the United States Department of Agriculture (USDA) textural soil classification system and the Unified Soil Classification System (USCS). Unfortunately, not all soil map units are classified according to the USDA or USCS systems, and previous attempts to provide a crosswalk table have been inconsistent. Random Forest machine learning model was used to create a USCS prediction model using USDA soil property variables. Important variables for predicting USCS code from available soil properties were USDA soil textures, percent organic material, and available water storage. Prediction error rates less than 2% were achieved compared to error rates of approximately 40% using crosswalk methods.