Geometrical learning, descriptive geometry, and biomimetic pattern recognition

Wang, Shoujue; Lai, Jun

doi:10.1016/j.neucom.2004.11.034

Cited by 34 publications

(10 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Video quality is decrease when translating in network, such as cable network or wireless network. From the theory of geometrical learning [7], some similar objects in high dimensional space are homeomorphism and can be measured as neighbors [8].…”

Section: Application Modelmentioning

confidence: 99%

A Method of Video Quality Assessment for CMMB Video Test Sequences

Zhu

Yang

2012

2012 International Conference on Computer Distributed Control and Intelligent Environmental Monitoring

View full text Add to dashboard Cite

Since CMMB is an important application in wireless communication field its video quality plays a critical role for widely use. In assessment of video sequence, classic method often applies algebra method in making a compute model, such as PSNR, which often result in difficult for alignment of video sequence numbers and leads to complex computation problems. In this paper, it presents a novel method bases in high dimensional space. The method considers an image as a point in high dimensional space. It uses special kernels to covering some image points from original image, then a noised image is put into space, the method is performed to find the most similar one in high-dimensional space and gives its similarity. Experimental results show that the propose method make the measurement easily and meet the real noised image sequence. The proposed method is constructive and this work provides a way for video quality assessment.

show abstract

Section: Application Modelmentioning

confidence: 99%

A Method of Video Quality Assessment for CMMB Video Test Sequences

Zhu

Yang

2012

2012 International Conference on Computer Distributed Control and Intelligent Environmental Monitoring

View full text Add to dashboard Cite

show abstract

“…A covering may be generated by a set of balls or ellipsoids following principal curve, for example using the piecewise linear skeletonization approximation to principal curves [95]. One algorithm of this type creates a "hypersausage" decision regions [96]. One-class SVM also provides covering in the kernel space [11].…”

Section: Transformation-based Meta-learningmentioning

confidence: 99%

Optimal Support Features for Meta-Learning

Duch

Maszczyk

Grochowski

2011

Studies in Computational Intelligence

View full text Add to dashboard Cite

Meta-learning has many aspects, but its final goal is to discover in an automatic way many interesting models for a given data. Our early attempts in this area involved heterogeneous learning systems combined with a complexity-guided search for optimal models, performed within the framework of (dis)similarity based methods to discover "knowledge granules". This approach, inspired by neurocognitive mechanisms of information processing in the brain, is generalized here to learning based on parallel chains of transformations that extract useful information granules and use it as additional features. Various types of transformations that generate hidden features are analyzed and methods to generate them are discussed. They include restricted random projections, optimization of these features using projection pursuit methods, similarity-based and general kernel-based features, conditionally defined features, features derived from partial successes of various learning algorithms, and using the whole learning models as new features. In the enhanced feature space the goal of learning is to create image of the input data that can be directly handled by relatively simple decision processes. The focus is on hierarchical methods for generation of information, starting from new support features that are discovered by different types of data models created on similar tasks and successively building more complex features on the enhanced feature spaces. Resulting algorithms facilitate deep learning, and also enable understanding of structures present in the data by visualization of the results of data transformations and by creating logical, fuzzy and prototype-based rules based on new features. Relations to various machine-learning approaches, comparison of results, and neurocognitive inspirations for meta-learning are discussed.

show abstract

“…A covering may be generated by a set of balls or ellipsoids following principal curve, for example using the piecewise linear skeletonization approximation to principal curves [104]. An algorithm of this type creating a "hypersausage" decision regions has been published recently [159]. More algorithms of this type should be developed, and their relations with neural algorithms investigated.…”

Section: Geometrical Perspectivementioning

confidence: 99%

Towards Comprehensive Foundations of Computational Intelligence

Duch

2007

Studies in Computational Intelligence

View full text Add to dashboard Cite

Abstract. Although computational intelligence (CI) covers a vast variety of different methods it still lacks an integrative theory. Several proposals for CI foundations are discussed: computing and cognition as compression, meta-learning as search in the space of data models, (dis)similarity based methods providing a framework for such meta-learning, and a more general approach based on chains of transformations. Many useful transformations that extract information from features are discussed. Heterogeneous adaptive systems are presented as particular example of transformation-based systems, and the goal of learning is redefined to facilitate creation of simpler data models. The need to understand data structures leads to techniques for logical and prototype-based rule extraction, and to generation of multiple alternative models, while the need to increase predictive power of adaptive models leads to committees of competent models. Learning from partial observations is a natural extension towards reasoning based on perceptions, and an approach to intuitive solving of such problems is presented. Throughout the paper neurocognitive inspirations are frequently used and are especially important in modeling of the higher cognitive functions. Promising directions such as liquid and laminar computing are identified and many open problems presented.

show abstract

Geometrical learning, descriptive geometry, and biomimetic pattern recognition

Cited by 34 publications

References 5 publications

A Method of Video Quality Assessment for CMMB Video Test Sequences

A Method of Video Quality Assessment for CMMB Video Test Sequences

Optimal Support Features for Meta-Learning

Towards Comprehensive Foundations of Computational Intelligence

Contact Info

Product

Resources

About