scholarly journals Greedy Convex Ensemble

Author(s):  
Thanh Tan Nguyen ◽  
Nan Ye ◽  
Peter Bartlett

We consider learning a convex combination of basis models, and present some new theoretical and empirical results that demonstrate the effectiveness of a greedy approach. Theoretically, we first consider whether we can use linear, instead of convex, combinations, and obtain generalization results similar to existing ones for learning from a convex hull. We obtain a negative result that even the linear hull of very simple basis functions can have unbounded capacity, and is thus prone to overfitting; on the other hand, convex hulls are still rich but have bounded capacities. Secondly, we obtain a generalization bound for a general class of Lipschitz loss functions. Empirically, we first discuss how a convex combination can be greedily learned with early stopping, and how a convex combination can be non-greedily learned when the number of basis models is known a priori. Our experiments suggest that the greedy scheme is competitive with or better than several baselines, including boosting and random forests. The greedy algorithm requires little effort in hyper-parameter tuning, and also seems able to adapt to the underlying complexity of the problem. Our code is available at https://github.com/tan1889/gce.

2016 ◽  
Vol 1 (3) ◽  
Author(s):  
Chih-Hung Chen ◽  
Ting-Ju Lin ◽  
Chih-Yu Chen

Based on the assumption that human behaviours are mainly affected by physical and animate environments, this empirical research takes the changeful and complex historical district in Tainan to observe wayfinding behaviours. An a priori analysis of the isovist fields is conducted to identify spatial characteristics. Three measures, the relative area, convexity, and circularity, are applied to scrutinize the possible stopping points, change of speed, and route choices. Accordingly, an experiment is carried out to observe spatial behaviours and different influences of social stimuli. Results show that social interactions afford groups and pairs to perform better than individual observers in wayfinding.© 2016. The Authors. Published for AMER ABRA by e-International Publishing House, Ltd., UK. Peer–review under responsibility of AMER (Association of Malaysian Environment-Behaviour Researchers), ABRA (Association of Behavioural Researchers on Asians) and cE-Bs (Centre for Environment-Behaviour Studies, Faculty of Architecture, Planning & Surveying, Universiti Teknologi MARA, MalaysiaKeywords: wayfinding; isovist; spatial perception and social stimuli; historic quarter


2013 ◽  
Vol 109 (5) ◽  
pp. 1259-1267 ◽  
Author(s):  
Devika Narain ◽  
Robert J. van Beers ◽  
Jeroen B. J. Smeets ◽  
Eli Brenner

In the course of its interaction with the world, the human nervous system must constantly estimate various variables in the surrounding environment. Past research indicates that environmental variables may be represented as probabilistic distributions of a priori information (priors). Priors for environmental variables that do not change much over time have been widely studied. Little is known, however, about how priors develop in environments with nonstationary statistics. We examine whether humans change their reliance on the prior based on recent changes in environmental variance. Through experimentation, we obtain an online estimate of the human sensorimotor prior (prediction) and then compare it to similar online predictions made by various nonadaptive and adaptive models. Simulations show that models that rapidly adapt to nonstationary components in the environments predict the stimuli better than models that do not take the changing statistics of the environment into consideration. We found that adaptive models best predict participants' responses in most cases. However, we find no support for the idea that this is a consequence of increased reliance on recent experience just after the occurrence of a systematic change in the environment.


2007 ◽  
Vol 7 (4) ◽  
pp. 321-340
Author(s):  
A. Masjukov

AbstractFor bivariate and trivariate interpolation we propose in this paper a set of integrable radial basis functions (RBFs). These RBFs are found as fundamental solutions of appropriate PDEs and they are optimal in a special sense. The condition number of the interpolation matrices as well as the order of convergence of the inter- polation are estimated. Moreover, the proposed RBFs provide smooth approximations and approximate fulfillment of the interpolation conditions. This property allows us to avoid the undecidable problem of choosing the right scale parameter for the RBFs. Instead we propose an iterative procedure in which a sequence of improving approx- imations is obtained by means of a decreasing sequence of scale parameters in an a priori given range. The paper provides a few clear examples of the advantage of the proposed interpolation method.


1970 ◽  
Vol 30 ◽  
pp. 19-31
Author(s):  
M Ashraf Uddin ◽  
M Matiar Rahman ◽  
M Saiful Islam Mallik

Generation of grid-scale (GS) and subgrid-scale (SGS) velocity fields is performed by direct filtering of DNS (Direct Numerical Simulation) data at a low Reynolds number in homogeneous isotropic turbulence in order to assess the spectral accuracy as well as the performance of filter functions for LES (Large Eddy Simulation). The filtering is performed using three classical filter functions: Gaussian, Tophat and Sharp cutoff filters and in all three cases the results are compared with three different filter widths for LES. Comparing the distributions of GS and SGS velocities, and the decay of turbulence with those from DNS fields through out the whole calculation we have found that among the three filter functions, the performance of Sharp cutoff filter is better than that of the other two filter functions in terms of both spatial spectra and the distribution of velocities. Furthermore, it is shown that the accuracy of the filtering approach does not depend only on the filter functions but also on the filter widths for LES. GANIT J. Bangladesh Math. Soc. (ISSN 1606-3694) 30 (2010) 19-31   DOI: http://dx.doi.org/10.3329/ganit.v30i0.8499


2020 ◽  
Vol 2020 ◽  
pp. 1-15
Author(s):  
Suiyuan Shen ◽  
Jinfa Xu

The internal uncertainty and external disturbance of the quadrotor will have a significant impact on flight control. Therefore, to improve the control system’s dynamic performance and robustness, the attitude active disturbance rejection controller (ADRC) of the quadrotor is established. Simultaneously, an adaptive genetic algorithm-particle swarm optimization (AGA-PSO) is used to optimize the controller parameters to solve the problem that the controller parameters are difficult to tune. The performance of the proposed ADRC is compared with that of the sliding mode controller (SMC). The simulations revealed that the dynamic performance and robustness of the ADRC is better than that of the SMC.


Glottotheory ◽  
2014 ◽  
Vol 5 (2) ◽  
Author(s):  
Ramon Ferrer-i-Cancho ◽  
Haitao Liu

AbstractMixing dependency lengths from sequences of different length is a common practice in language research. However, the empirical distribution of dependency lengths of sentences of the same length differs from that of sentences of varying length. The distribution of dependency lengths depends on sentence length for real sentences and also under the null hypothesis that dependencies connect vertices located in random positions of the sequence. This suggests that certain results, such as the distribution of syntactic dependency lengths mixing dependencies from sentences of varying length, could be a mere consequence of that mixing. Furthermore, differences in the global averages of dependency length (mixing lengths from sentences of varying length) for two different languages do not simply imply a priori that one language optimizes dependency lengths better than the other because those differences could be due to differences in the distribution of sentence lengths and other factors.


Physics ◽  
2021 ◽  
Vol 3 (3) ◽  
pp. 569-578
Author(s):  
Spyridon Koutsoumpos ◽  
Panagiotis Giannios ◽  
Konstantinos Moutzouris

Critical angle refractometry is an established technique for determining the refractive index of liquids and solids. For transparent samples, the critical angle refractometry precision is limited by incidence angle resolution. For lossy samples, the precision is also affected by reflectance measurement error. In the present study, it is demonstarted that reflectance error can be practically eliminated, provided that the sample’s extinction coefficient is a priori known with sufficient accuracy (typically, better than 5%) through an independent measurement. Then, critical angle refractometry can be as precise with lossy media as with transparent ones.


2019 ◽  
Vol 8 (1) ◽  
pp. 80-93
Author(s):  
Kourosh Parand ◽  
Yasaman Lotfi ◽  
Jamal Amani Rad

AbstractIn this paper the boundary layer flow of a micro-polar fluid due to a linearly stretching sheet which is a non-linear system two-point boundary value problem (BVP) on semi-infinite interval has been considered. This the sheets are included the suction and injection. We solve this problem by two different collecation approaches and compare their results with solution of other methods. The proposed approaches are equipped by the direct (DRBF) and indirect radial basis functions (IRBF). Direct approach (DRBF) is based on a differential process and indirect approach (IRBF) is based on an integration process. These methods reduce solution of the problem to solution of a system of algebraic equations. Numerical results and residual norm show that the IRBF performs better than the common DRBF, and has an acceptable accuracy and high rate of convergence of IRBF process.


Sign in / Sign up

Export Citation Format

Share Document