Information-Theoretic Active SOM for Improving Generalization Performance
Journal Title: International Journal of Advanced Research in Artificial Intelligence(IJARAI) - Year 2016, Vol 5, Issue 8
Abstract
In this paper, we introduce a new type of information-theoretic method called “information-theoretic active SOM”, based on the self-organizing maps (SOM) for training multi-layered neural networks. The SOM is one of the most important techniques in unsupervised learning. However, SOM knowledge is sometimes ambiguous and cannot be easily interpreted. Thus, we introduce the information-theoretic method to produce clearer and interpretable representations. The present method extends this information-theoretic approach into supervised learning. The main contribution can be summarized by three points. First, it is shown that clear representations by the information-theoretic method can be effective in training supervised learning. Second, the method is sufficiently simple where there are two separated components, namely, information maximization and error minimization component. Usually, two components are mixed in one framework, and it is difficult to compromise between them. In addition, the knowledge obtained by this information-theoretic SOM can be used to solve the shortage of unlabeled data, because the information maximization component is unsupervised and can process all input data with and without labels. The method was applied to the well-known image segmentation datasets. Experimental results showed that clear weights were produced and generalization performance was improved by using the information-theoretic SOM. In addition, the final results were stable, almost independent of the parameter values.
Authors and Affiliations
Ryotaro Kamimura
Keywords
Related Articles
- EP ID EP101698
- DOI 10.14569/IJARAI.2016.050804
- Views 113
- Downloads 0
System for Human Detection in Image Based on Intel Galileo
The aim of this paper is a comparative analysis of methods for motion detection and human recognition in the image. Authors propose the own solution following the comparative analysis of current approaches. Then au...
Appropriate Tealeaf Harvest Timing Determination Referring Fiber Content in Tealeaf Derived from Ground based Nir Camera Images
Method for most appropriate tealeaves harvest timing with the reference to the fiber content in tealeaves which can be estimated with ground based Near Infrared (NIR) camera images is proposed. In the proposed meth...
3D Map Creation Based on Knowledgebase System for Texture Mapping Together with Height Estimation Using Objects’ Shadows with High Spatial Resolution Remote Sensing Satellite Imagery Data
Method for 3D map creation based on knowledgebase system for texture mapping together with height estimation using objects’ shadows with high spatial resolution of remote sensing satellite imagery data is proposed....
Blurring and Deblurring Digital Images Using the Dihedral Group
A new method of blurring and deblurring digital images is presented. The approach is based on using new filters generating from average filter and H-filters using the action of the dihedral group. These filters are...
Using Mining Predict Relationships on the Social Media Network: Facebook (FB)
The objective of this paper is to study on the most famous social networking site Facebook and other online social media networks (OSMNs) based on the notion of relationship or friendship. This paper discussed the...