Contradiction Resolution between Self and Outer Evaluation for Supervised Multi-Layered Neural Networks
Journal Title: International Journal of Advanced Research in Artificial Intelligence(IJARAI) - Year 2013, Vol 2, Issue 7
Abstract
In this paper, we propose a new type of informationtheoretic method. We suppose that a neuron should be evaluated from different points of view to precisely discern its properties. In this paper, we restrict ourselves to two types of evaluation methods for neurons, namely, self and outer-evaluation. A neuron fires only as a result of evaluating itself, while the neuron can fire as a result of evaluation by all surrounding neurons. Selfand outer-evaluation should be equivalent to each other. When contradiction between two types of evaluation exists, the contradiction should be as small as possible. Contradiction between self- and outer-evaluations is realized in terms of the Kullback- Leibler divergence between two types of neurons. Contradiction between self- and outer-evaluation can be resolved by decreasing the contradiction ratio between the two types of evaluation in terms of KL divergence. This method is expected to extract the main features in input patterns, if those are shared by two types of evaluation. We applied the method to two data sets, namely, the logistic and dollar-yen exchange rate data. In both problems, experimental results showed that visualization performance could be improved, leading to clearer class structure for both problems. In addition, when visualization was improved, generalization performance did not necessarily degrade, showing the possibility of networks with better visualization and prediction performance.
Authors and Affiliations
Ryotaro Kamimura
Trend Analysis of Relatively Large Diatoms Which Appear in the Intensive Study Area of the Ariake Sea, Japan in Winter (2011-2015) based on Remote Sensing Satellite Data
Behavior of relatively large size of diatoms which appear in the Ariake Sea areas, Japan in winter based on remote sensing satellite data is clarified. Through experiments with Terra and AQUA MODIS data derived chl...
A Rank Aggregation Algorithm for Ensemble of Multiple Feature Selection Techniques in Credit Risk Evaluation
In credit risk evaluation the accuracy of a classifier is very significant for classifying the high-risk loan applicants correctly. Feature selection is one way of improving the accuracy of a classifier. It provide...
Direction for Artificial Intelligence to Achieve Sapiency Inspired by Homo Sapiens
Artificial intelligence technology has developed significantly in the past decades. Although many computational programs are able to approximate many cognitive abilities of Homo sapiens, the intelligence and sapien...
FlexRFID: A Security and Service Control Policy-Based Middleware for Context-Aware Pervasive Computing
Ubiquitous computing targets the provision of seamless services and applications by providing an environment that involves a variety of devices having different capabilities. The design of applications in these env...
Accurate Topological Measures for Rough Sets
Data granulation is considered a good tool of decision making in various types of real life applications. The basic ideas of data granulation have appeared in many fields, such as interval analysis, quantization, r...