Search results for "generalization"
showing 10 items of 250 documents
Adaptive interpolation with maximum order close to discontinuities
2022
Abstract Adaptive rational interpolation has been designed in the context of image processing as a new nonlinear technique that avoids the Gibbs phenomenon when we approximate a discontinuous function. In this work, we present a generalization to this method giving explicit expressions for all the weights for any order of the algorithm. It has a similar behavior to weighted essentially non oscillatory (WENO) technique but the design of the weights in this case is more simple. Also, we propose a new way to construct them obtaining the maximum order near the discontinuities. Some experiments are performed to demonstrate our results and to compare them with standard methods.
2014
This paper investigates the proficiency of support vector machine (SVM) using datasets generated by Tennessee Eastman process simulation for fault detection. Due to its excellent performance in generalization, the classification performance of SVM is satisfactory. SVM algorithm combined with kernel function has the nonlinear attribute and can better handle the case where samples and attributes are massive. In addition, with forehand optimizing the parameters using the cross-validation technique, SVM can produce high accuracy in fault detection. Therefore, there is no need to deal with original data or refer to other algorithms, making the classification problem simple to handle. In order to…
Three Wives Problem and Shapley Value
2014
We examine the Talmudic three wives problem, which is a generalization of the Talmudic contested garment problem solved by Aumann and Maschler (1985) using coalitional procedure. This problem has many practical applications. In an attempt to unify all Talmudic methods, Guiasu (2010, 2011) asserts that it can be explained in terms of “run-to-the-bank”, that is, of Shapley value in a “cumulative game”. It can be challenged because the coalitional procedure yields the same result as the nucleolus, which corresponds to a “dual game”. As Guiasu’s solution is paradoxical (it has all the appearances of truth), my contribution consists in explaining the concepts, particularly truncation, that play …
Do we need algebraic-like computations? A reply to Bonatti, Pena, Nespor, and Mehler (2006).
2006
L. L. Bonatti, M. Pena, M. Nespor, and J. Mehler (2006) argued that P. Perruchet, M. D. Tyler, N. Galland, and R. Peereman (2004) confused the notions of segmentation and generalization by ignoring the evidence for generalization in M. Pena, L. L. Bonatti, M. Nespor, and J. Mehler (2002). In this reply, the authors reformulate and complement their initial arguments, showing that their way of dealing with segmentation and generalization is not due to confusion or ignorance but rather to the fact that the tests used in Pena et al. make it likely that neither segmentation nor generalization were captured in their experiments. Finally, the authors address the challenge posed by Pena et al. of a…
Order statistics-based parametric classification for multi-dimensional distributions
2013
Traditionally, in the field of Pattern Recognition (PR), the moments of the class-conditional densities of the respective classes have been used to perform classification. However, the use of phenomena that utilized the properties of the Order Statistics (OS) were not reported. Recently, in [10,8], we proposed a new paradigm named CMOS, Classification by the Moments of Order Statistics, which specifically used these quantifiers. It is fascinating that CMOS is essentially ''anti''-Bayesian in its nature because the classification is performed in a counter-intuitive manner, i.e., by comparing the testing sample to a few samples distant from the mean, as opposed to the Bayesian approach in whi…
Non-Local Scattering Kernel and the Hydrodynamic Limit
2007
In this paper we study the interaction of a fluid with a wall in the framework of the kinetic theory. We consider the possibility that the fluid molecules can penetrate the wall to be reflected by the inner layers of the wall. This results in a scattering kernel which is a non-local generalization of the classical Maxwell scattering kernel. The proposed scattering kernel satisfies a global mass conservation law and a generalized reciprocity relation. We study the hydrodynamic limit performing a Knudsen layer analysis, and derive a new class of (weakly) nonlocal boundary conditions to be imposed to the Navier-Stokes equations.
Approximation of exit times for one-dimensional linear diffusion processes
2020
International audience; In order to approximate the exit time of a one-dimensional diffusion process, we propose an algorithm based on a random walk. Such an algorithm was already introduced in both the Brownian context and the Ornstein-Uhlenbeck context, that is for particular time-homogeneous diffusion processes. Here the aim is therefore to generalize this efficient numerical approach in order to obtain an approximation of both the exit time and position for a general linear diffusion. The main challenge of such a generalization is to handle with time-inhomogeneous diffusions. The efficiency of the method is described with particular care through theoretical results and numerical example…
A non-linear stochastic approach of ligaments and tendons fractional-order hereditariness
2020
Abstract In this study the non-linear hereditariness of knee tendons and ligaments is framed in the context of stochastic mechanics. Without losing the possibility of generalization, this work was focused on knee Anterior Cruciate Ligament (ACL) and the tendons used in its surgical reconstruction. The proposed constitutive equations of fibrous tissues involves three material parameters for the creep tests and three material parameters for relaxation tests. One-to-one relations among material parameters estimated in creep and relaxations were established and reported in the paper. Data scattering, observed with a novel experimental protocol used to characterize the mechanics of the tissue, w…
Interpretable Option Discovery Using Deep Q-Learning and Variational Autoencoders
2021
Deep Reinforcement Learning (RL) is unquestionably a robust framework to train autonomous agents in a wide variety of disciplines. However, traditional deep and shallow model-free RL algorithms suffer from low sample efficiency and inadequate generalization for sparse state spaces. The options framework with temporal abstractions [18] is perhaps the most promising method to solve these problems, but it still has noticeable shortcomings. It only guarantees local convergence, and it is challenging to automate initiation and termination conditions, which in practice are commonly hand-crafted.