Nonparametric Function Fitting in the Presence of Nonstationary Noise

Abstract. In the paper, the parallel realization of the Boltzmann Restricted Machine (RBM) is proposed. The implementation intends to use multicore architectures of modern CPUs and Intel Xeon Phi coprocessor. The learning procedure is based on the matrix description of RBM, where the learning samples are grouped into packages, and represented as matrices. The influence of the package size on convergence of learning, as well as on performance of computation, are studied for various number of threads, using conventional CPU and Intel Phi architecures. Our research confirms a potential usefulness of MIC parallel architecture for implementation of RBM and similar algorithms.

Section: Discussionmentioning

confidence: 99%

Adaptation of RBM Learning for Intel MIC Architecture

Olas

Mleczko

Nowicki

et al. 2015

“…It is worth to note that many computational intelligence methods (see e.g. [1,11,17,18,19,21,68,49,64,65]) are succesfully used in pattern recognition (see e.g. [20,22,23,55,56,57,58]) and modelling (see e.g.…”

Section: Introductionmentioning

confidence: 99%

New Fast Algorithm for the Dynamic Signature Verification Using Global Features Values

Zalasiński

Cpałka

Hayashi

2015

Abstract. Identity verification based on the dynamic signature is an important issue of biometrics. There are many effective methods to the signature verification which take into account the dynamic characteristics of the signature (e.g. velocity of the pen, the pen's pressure on the surface of the graphic tablet, etc.). Among these methods, the ones based on the so-called global features are very important. In our previous paper we have proposed new algorithm for evolutionary selection of the dynamic signature global features, which selects a subset of features individually for each user. Algorithm proposed in this paper is a faster version of the method proposed earlier. During development of the algorithm we resigned from using evolutionary selection of global features and standardized working of the classifier in the context of all users. The paper contains the simulation results for the BioSecure database of the dynamic signatures.

“…[16,69,83,97,104,105,106,107]) and many other (see e.g. [32,33,34,50]) issues. This paper is presented as follows: in Section 2 a description of proposed system and its tuning process for nonlinear classification is presented.…”

Section: Introductionmentioning

confidence: 99%

A New Interpretability Criteria for Neuro-Fuzzy Systems for Nonlinear Classification

Łapa

Cpałka

Galushkin

2015

Abstract. In this paper a new approach for construction of neuro-fuzzy systems for nonlinear classification is introduced. In particular, we concentrate on the flexible neuro-fuzzy systems which allow us to extend notation of rules with weights of fuzzy sets. The proposed approach uses possibilities of hybrid evolutionary algorithm and interpretability criteria of expert knowledge. These criteria include not only complexity of the system, but also semantics of the rules. The approach presented in our paper was tested on typical nonlinear classification simulation problems.