A new intelligent pattern classifier based on deep-thinking
- PDF / 3,133,579 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 92 Downloads / 159 Views
(0123456789().,-volV)(0123456789(). ,- volV)
EXTREME LEARNING MACHINE AND DEEP LEARNING NETWORKS
A new intelligent pattern classifier based on deep-thinking Zhenyi Shen1
•
Zhihong Man1 • Zhenwei Cao1 • Jinchuan Zheng1
Received: 30 December 2018 / Accepted: 29 August 2019 Ó Springer-Verlag London Ltd., part of Springer Nature 2019
Abstract A new intelligent pattern classifier based on the human being’s thinking logics is developed in this paper, aiming to approximate the optimal design process and avoid the matrix inverse computation in conventional classifier designs. It is seen that the proposed classifier has no parameters to be determined via mathematical optimization. Instead, it is built by using the correlation principles to construct the clusters at first. The middle-level feature vectors can then be extracted from the statistical information of the correlations between the input data and the ones in each pattern cluster. For accurate classification purpose, the advanced feature vectors are generated with the moments’ information of the middle-level feature vectors. After that, Bayesian inference is implemented to make decisions from the weighted sum of the advanced feature components. In addition, a real-time fine-tuning loop (layer) is designed to adaptively ‘‘widen’’ the border of each pattern clustering region such that the input data can be directly classified once they are located in one of the clustering regions. An experiment for the classification of the handwritten digit images from the MNIST database is performed to show the excellent performance and effectiveness of the proposed intelligent pattern classifier. Keywords Deep-thinking pattern classifier Bayesian inference Unsupervised learning Correlation principle
1 Introduction The human brain has a complicated structure to support powerful mental activities. In the last few decades, the brain scientists have discovered that the lobes in our brain play different roles in human information pre-processing, memory storage, learning, logical reasoning and problemsolving processes [1]. It is seen that with these lobes working in parallel and cooperating with each other, human beings are able to effectively learn the dynamic characteristics of both labelled and unlabelled data, and achieve
& Zhihong Man [email protected] Zhenyi Shen [email protected] Zhenwei Cao [email protected] Jinchuan Zheng [email protected] 1
Faculty of Science, Engineering and Technology, Swinburne University of Technology, Hawthorn, Melbourne, VIC 3122, Australia
high classification accuracy with a small training dataset. Human beings’ thinking activities are actually based on the parallel processing and logical reasoning, which are often regarded as the deep-thinking for distinguishing from animals. Motivated by human being’s deep-thinking mechanism, many classification techniques developed through finding out data features and then performing the pattern classification in feature space. Recently, the researchers have combined conventional pattern classification
Data Loading...
