Given any generative classifier based on an inexact density model, we can define a discriminative counterpart that reduces its asymptotic error rate, while increasing the estimation variance. An optimal bias-variance ...
详细信息
ISBN:
(纸本)9780769530697
Given any generative classifier based on an inexact density model, we can define a discriminative counterpart that reduces its asymptotic error rate, while increasing the estimation variance. An optimal bias-variance balance might be found using Hybrid Generative-Discriminative (HGD) approaches. In these paper these methods are defined in a unfied framework. this allow us to find sufficient conditions under which an improvement in generalization performances is guaranteed Numerical experiments illustrate the well fondness of our statements.
Developing intelligent tools to extract information from data collections has long been of critical importance in fields such as knowledge discovery, information retrieval, patternrecognition, and databases. Withthe...
详细信息
ISBN:
(纸本)9780769530697
Developing intelligent tools to extract information from data collections has long been of critical importance in fields such as knowledge discovery, information retrieval, patternrecognition, and databases. Withthe advent of electronic medical records and medical data repositories there is new potential to apply these techniques to the analysis of biomedical data sets. Looking for complex patterns within large biomedical data repositories and discovering previously unexpected associations can be of particular interest for understanding the physiology and functionality of the human body as well as tracing the roots of diseases. In the context of a research hospital these analyses may lead to further directed research, better diagnostic capabilities, and improved patient outcomes. this paper describes an implementation of a knowledge discovery algorithm aimed at such data sets.
In traditional flat neural network, the topologic configurations are needed to be rebuilt withthe width of cold strip changing. So that, the large learn assignment, slow convergence and local minimal in the network a...
详细信息
ISBN:
(纸本)9781424409723
In traditional flat neural network, the topologic configurations are needed to be rebuilt withthe width of cold strip changing. So that, the large learn assignment, slow convergence and local minimal in the network are observed. Moreover, the structure of the traditional neural network according to the experience has been proved that the model is time-consuming and complex. lit, this paper, a new approach of flatness patternrecognition is proposed based on the CMAC neural network. the difference of fuzzy distances between samples and the basic patterns is introduced as the inputs of the CMAC network. Simultaneity momentum term is imported to update the weight of this neural network. the new approach withthe advantages, such as fast learning speed, good generalization, and easiness to implement, is efficient and intelligent. the simulation results show that the speed and accuracy of the flat patternrecognition model are improved obviously.
the proceedings contain 71 papers. the topics discussed include: language understanding and unified cognitive science;cognitive informatics foundations of nature and machine intelligence;challenges in the design of ad...
ISBN:
(纸本)1424413273
the proceedings contain 71 papers. the topics discussed include: language understanding and unified cognitive science;cognitive informatics foundations of nature and machine intelligence;challenges in the design of adaptive, intelligent, and cognitive systems;a approach to representation changes while executing problem solver intelligent systems;formal linguistics and the deductive grammar;towards a spatial representation for the meta cognitive process layer of cognitive informatics;the visual implications of inspection time;image decomposition and reconstruction using two-dimensional complex-valued Gabor wavelets;cognitive informatics in automatic pattern understanding;a cognitive data visualization method based on hyper surface;and a simple high accuracy approach for face recognition.
the proceedings contain 363 papers. the topics discussed include: design and implementation of a UWB peer-to-peer communication network;an improved feedback neural network for the design of all-pass phase equalizers;M...
详细信息
ISBN:
(纸本)1424409837
the proceedings contain 363 papers. the topics discussed include: design and implementation of a UWB peer-to-peer communication network;an improved feedback neural network for the design of all-pass phase equalizers;MRF-based adaptive approach for foreground segmentation under sudden illumination change;a generic interface methodology for bridging application systems and speech recognizers;internetworking security model based on intelligent system;detection of ventricular Arrhythmias using roots location in AR-modelling;comparison of spectral subtraction methods used in noise suppression algorithms;nonstationary power signal processing and patternrecognition using genetic algorithm;engineering polymeric optical fibers with desired properties;data and decision fusion for distributed spectrum sensing in cognitive radio networks;and comparative study of input-balanced realization-based digital filter structures.
the highest fidelity representations of realistic real-world materials currently comprise Bidirectional Texture Functions (BTF). the BTF is a six-dimensional function depending on view and illumination directions as w...
详细信息
the highest fidelity representations of realistic real-world materials currently comprise Bidirectional Texture Functions (BTF). the BTF is a six-dimensional function depending on view and illumination directions as well as on planar texture coordinates. the huge size of such measurements, typically in the form of thousands of images covering all possible combinations of illumination and viewing angles, has prohibited their practical exploitation, and obviously some compression and modelling method of these enormous BTF data spaces is inevitable. the two proposed approaches combine BTF spatial clustering with cluster index modelling by means of efficient Markov random field models. the methods allow the generation of a seamless cluster index of arbitrary size to cover large virtual 3D object surfaces. Both methods represent original BTF data using a set of local spatially dependent Bidirectional Reflectance Distribution Function (BRDF) values which are combined according to the synthesized cluster index and illumination/viewing directions by means of two types of Markov random field models. BTF data compression using both methods is about 1: 200 and their synthesis is very fast.
Krawtchouk moments are not rotational invariant. Hence, this paper introduces a new radial Krawtchouk moments using polar representation of an image by a one dimensional orthogonal discrete weighted Krawtchouk polynom...
详细信息
the ability to associate images is the basis for learning relationships involving vision, hearing, tactile sensation, and kinetic motion. A new architecture is described that has only local, recurrent connections, but...
详细信息
ISBN:
(纸本)9781424413270
the ability to associate images is the basis for learning relationships involving vision, hearing, tactile sensation, and kinetic motion. A new architecture is described that has only local, recurrent connections, but can directly form global image associations. this architecture has many similarities to the structure Of the neocortex, including the division into Brodmann areas, the distinct internal and external lamina, and the pattern of neuron interconnection. Analogous to the bits in an SR flip-flop, two arbitrary images can hold each other in place in an association processor and thereby form a short-term image memory. Overlay masks can focus attention on specific image regions. Spherically symmetric wavelets, identical to those found in the receptive fields of the retina, enable efficient image computations. Stability and noise reduction in reciprocal continuous wavelet transform representations can be achieved using an orthogonal projection based on the reproducing kernel.
the proceedings contain 76 papers. the topics discussed include: enabling advanced and context-dependent access control in RDF stores;automatically composing data workflows with relational descriptions and shim servic...
详细信息
ISBN:
(纸本)3540762973
the proceedings contain 76 papers. the topics discussed include: enabling advanced and context-dependent access control in RDF stores;automatically composing data workflows with relational descriptions and shim services;how service choreography statistics reduce the ontology mapping problem;kernel methods for mining instance data in ontologies;an ontology design pattern for representing relevance in OWL;scalable cleanup of information extraction data using ontologies;a cognitive support framework for ontology mapping;discovering simple mappings between relational database schemas and ontologies;an empirical study of instance-based ontology matching;conjunctive queries for a tractable fragment of OWL 1.1;ontology-based controlled natural language editor using CFG with lexical dependency;and web search personalization via social bookmarking and tagging.
the recognition of hand-written Chinese characters using Mahalanobis distance is extensively utilized in bank cheque processing applications. the Mahalanobis distance, defined by the innovation and its covariance, is ...
详细信息
ISBN:
(纸本)9781424409723
the recognition of hand-written Chinese characters using Mahalanobis distance is extensively utilized in bank cheque processing applications. the Mahalanobis distance, defined by the innovation and its covariance, is compared among several target character classes, and the computation is a time-consuming operation. this paper presents an efficient computation for this process. the method described here can be summarized as an incremental, non-decreasing computation for the Mahalanobis distance;if the incrementally computed value exceeds the threshold then the computation is stopped. the elements of covariance and innovation are only computed if they are used, and progressivity is the major advantage of the method. this method is based upon the square-root-free Cholesky's factorization. Experiment shows the method proposed here is effective in financial hand-written character recognition.
暂无评论