Representation Of Input Patterns Research Articles

Classification of NPPs transients using change of representation technique: A hybrid of unsupervised MSOM and supervised SVM

This study introduces a new identifier for nuclear power plants (NPPs) transients. The proposed identifier changes the representation of input patterns. Change of representation is a semi-supervised learning algorithm which employs both of labeled and unlabeled input data. In the first step, modified self-organizing map (MSOM) carries out an unsupervised learning algorithm on labeled and unlabeled patterns and generates a new metric for input data. In the second step, support vector machine (SVM) as a supervised learning algorithm classifies the input patterns using the generated metric of the first step. In contrast to unsupervised learning algorithms, the proposed identifier does not discard the collected information. The proposed identifier is examined by the Iris flower dataset and the Bushehr nuclear power plant (BNPP) transients. In overall, results show good performance of the developed identifier. Training with small fraction of labeled patterns, classification only by the sign of the classifiers outputs, and modular identification are main advantages of the proposed identifier. Results indicate that the developed MSOM is not able to cluster correctly quasi-static transients such as uncontrolled withdrawal of control rods (UWCR) in presence of steady state patterns. Quasi-static transients are very similar to steady state. Moreover, imposed noise on input data may eliminate minor differences among the patterns of them. This may cause wrong training of SVM. This is the most important challenge of semi-supervised learning. In other words, if the knowledge on density of unlabeled input patterns do not carry useful information for prediction of the targets of unlabeled patterns, then semi-supervised learning may degrade the prediction accuracy. Therefore, the proposed identifier is more appropriate for classification of transients in which either clusters have been apart or small noises have been imposed on data.

Adaptive Synaptogenesis Constructs Neural Codes That Benefit Discrimination

Intelligent organisms face a variety of tasks requiring the acquisition of expertise within a specific domain, including the ability to discriminate between a large number of similar patterns. From an energy-efficiency perspective, effective discrimination requires a prudent allocation of neural resources with more frequent patterns and their variants being represented with greater precision. In this work, we demonstrate a biologically plausible means of constructing a single-layer neural network that adaptively (i.e., without supervision) meets this criterion. Specifically, the adaptive algorithm includes synaptogenesis, synaptic shedding, and bi-directional synaptic weight modification to produce a network with outputs (i.e. neural codes) that represent input patterns proportional to the frequency of related patterns. In addition to pattern frequency, the correlational structure of the input environment also affects allocation of neural resources. The combined synaptic modification mechanisms provide an explanation of neuron allocation in the case of self-taught experts.

On hardness of several string indexing problems

Let D={d1,d2,…,dD} be a collection of D string documents of n characters in total. The two-pattern matching problems ask to index D for answering the following queries efficiently.•Report/count the unique documents containing P1 and P2.•Report/count the unique documents containing P1, but not P2.Here P1 and P2 represent input patterns of length p1 and p2 respectively. Linear space data structures with O(p1+p2+nklogO(1)⁡n) query cost are already known for the reporting version, where k represents the output size. For the counting version (i.e., report the value k), a simple linear-space index with O(p1+p2+n) query cost can be constructed in O(n3/2) time. However, it is still not known if these are the best possible bounds for these problems. In this paper, we show a strong connection between these string indexing problems and the boolean matrix multiplication problem. Based on this, we argue that these results cannot be improved significantly using purely combinatorial techniques. We also provide an improved upper bound for a related problem known as common colors query problem.

MAROR: Multi-Level Abstraction of Association Rule Using Ontology and Rule Schema

Many large organizations have multiple databases distributed over different branches. Number of such organizations is increasing over time. Thus, it is necessary to study data mining on multiple databases. Most multi-databases mining (MDBM) algorithms for association rules typically represent input patterns at a single level of abstraction. However, in many applications of association rules - e.g., Industrial discovery, users often need to explore a data set at multiple levels of abstraction, and from different points of view. Each point of view corresponds to set of beliefs (and representational) commitments regarding the domain of interest. Using domain ontologies, we strengthen the integration of user knowledge in the mining and post-processing task. Furthermore, an interactive and iterative framework is designed to assist the user along the analyzing task at different levels. This paper formalizes the problem of association rules using ontologies in multi-database mining, describes an ontology-driven association rules algorithm to discoverer rules at multiple levels of abstraction and presents preliminary results in petroleum field to demonstrate the feasibility and applicability of this proposed approach.

Similarity interaction in information-theoretic self-organizing maps

In this paper, we propose a new information-theoretic computational method called ‘similarity interaction’ for improving visualization. Due to the fixed arrangement of neurons in the self-organizing maps, similarity between neurons is not necessarily a faithful representation of the actual similarity between neurons. To relax the fixed arrangement, we introduce a method called ‘similarity interaction’, because we integrate the information of connection weights into that of neurons. We applied our method to three problems, namely teaching assistant evaluation, automobile data, and dermatology data. In all three problems, we succeeded in demonstrating the better performance of our method through visual inspection and quantitative evaluation. Our method is the first step towards the interaction of multiple components in a neural network for finer representations of input patterns.

Comparison of graph-based methods for non-linear dimensionality reduction

In this paper, four broadly representative graph-based techniques for manifold learning namely Isomap, Maximum Variance Unfolding (MVU), locally linear embedding and Laplacian eigenmaps have been reviewed and compared for non-linear dimensionality reduction. These methods begin by constructing a sparse graph in which the nodes represent input patterns and the edges represent neighbourhood relations. From these graphs, matrices can be constructed whose spectral decompositions reveal the low dimensional structure of the submanifold. All the four techniques are implemented on Swiss roll, helix, twin peak and broken Swiss roll dataset.

Improving feature extraction performance of greedy network-growing algorithm by inverse euclidean distance

We propose here a new computational method for the information-theoretic method, called the greedy network-growing algorithm, to facilitate a process of information acquisition. We have so far used the sigmoidal activation function for competitive unit outputs. The method can effectively suppress many competitive units by generating strongly negative connections. However, because methods with the sigmoidal activation function are not very sensitive to input patterns, we have observed that in some cases final representations obtained by the method do not necessarily faithfully describe input patterns. To remedy this shortcoming, we employ the inverse of distance between input patterns and connection weights for competitive unit outputs. As the distance becomes smaller, competitive units are more strongly activated. Thus, winning units tend to represent input patterns more faithfully than in the previous method with the sigmoidal activation function. We applied the new method to artificial data analysis and animal classification. Experimental results confirmed that more information can be acquired and more explicit features can be extracted by our new method.

Separability of internal representations in multilayer perceptrons with application to learning

It is mathematically investigated as to what kind of internal representations are separable by single output units of three-layer perceptrons. A topological description is given for the necessary and sufficient condition that hidden layer representations of input patterns are separable by the output unit. An efficient algorithm is proposed for checking whether or not a hidden layer representation is linearly separable and, if not, for specifying inseparable portions in the partition. Application of the algorithm to learning of three-layer perceptrons is presented in which redundant units are utilized to reduce inseparable partition into separable one. Polynomial learnability from examples and queries is shown for the proposed learning algorithm.