Seeking classifier models that are not overconfident and that better represent the inherent uncertainty over a set of choices, we extend an objective for semi-supervised learning for neural networks to two models from the ratio semi-definite classifier (RSC) family. We show that the RSC family of classifiers produces smoother transitions between classes on a vowel classification task, and that ...

This paper presents a novel classifier based on collaborative representation and multiple onedimensional embedding with applications to face recognition. To use multiple 1-D embedding (1DME) framework in semi-supervised learning is first proposed by one of the authors, J. Wang, in 2014. The main idea of the multiple 1-D embedding is the following: Given a high-dimensional data set, we first map...

Most methods in machine learning are described as either discriminative or generative. The former often attain higher predictive accuracy, while the latter are more strongly regularized and can deal with missing data. Here, we propose a new framework to combine a broad class of discriminative and generative models, interpolating between the two extremes with a multiconditional likelihood object...

The interpolation, prediction, and feature analysis of fine-gained air quality are three important topics in the area of urban air computing. The solutions to these topics can provide extremely useful information to support air pollution control, and consequently generate great societal and technical impacts. Most of the existing work solves the three problems separately by different models. In...

This article aims at giving a simplified presentation of a new adaptive semi-Lagrangian scheme for solving the (1 + 1)dimensional Vlasov-Poisson system, which was developed in 2005 with Michel Mehrenberger and first described in (Campos Pinto and Mehrenberger, 2007). The main steps of the analysis are also given, which yield the first error estimate for an adaptive scheme in the context of the ...

We propose a Semi-Lagrangian scheme coupled with Radial Basis Function interpolation for approximating a curvature-related level set model, which has been proposed by Zhao et al. in [19] to reconstruct unknown surfaces from sparse, possibly noisy data sets. The main advantages of the proposed scheme are the possibility to solve the level set method on unstructured grids, as well as to concentra...

In this paper we propose a novel general framework for unsupervised model adaptation. Our method is based on entropy which has been used previously as a regularizer in semi-supervised learning. This technique includes another term which measures the stability of posteriors w.r.t model parameters, in addition to conditional entropy. The idea is to use parameters which result in both low conditio...

We study the cohomology of invertible sheaves C on surfaces X, blowings-up of P2 at points pl,...,p, in general position (generic rational surfaces). The main theme is when such sheaves have the natural cohomology, i.e. at most one cohomology group is non zero. Our approach is geometrical. On one hand we are lead to deform the configuration of points to a special position, notably surfaces with...

In this paper we propose a new prism for studying deep learning motivated by connections between deep learning and evolution. Our main contributions are: • We introduce of a sequence of increasingly complex hierarchical generative models which interpolate between standard Markov models on trees (phylogenetic models) and deep learning models. • Formal definitions of classes of algorithms that ar...

We introduce a semiempirical method to correct the systematic equilibrium lattice parameters underestimation present in first principles calculations based on the local density approximation. The method consists in performing calculations under a negative pressure such that the calculated equilibrium volume matches the experimentally observed one. We find that elastic properties obtained under ...