Evolutionary Hierarchical Sparse Extreme Learning Autoencoder Network for Object Recognition

Zeng, Yujun; Qian, Lilin; Ren, Junkai

doi:10.3390/sym10100474

Cited by 6 publications

(8 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Inspired by polynomial approximation of non-linear functions, we use a multipath and piecewise linear approximation in the design of the fast sparse DNN (FSDNN) network, thus combining the hidden layer weight matrix and biased random sampling with sparsity, and further design the corresponding low-time-consuming iterative optimization algorithm. In addition, the proposed FSDNN in this paper is different from the previous models [28]- [30] in terms of the core modules of the network, framework design and optimization algorithm. The main contributions of this study can be summarized as follows.…”

Section: Introductionmentioning

confidence: 89%

Fast Sparse Deep Neural Networks: Theory and Performance Analysis

Zhao

Jiao

2019

IEEE Access

View full text Add to dashboard Cite

In this paper, fast sparse deep neural networks that aim to offer an alternative way of learning in a deep structure are proposed. We examine some optimization algorithms for traditional deep neural networks and find that deep neural networks suffer from a time-consuming training process because of a large number of connecting parameters in layers and layers. To reduce time consumption, we propose fast sparse deep neural networks, which mainly consider the following two aspects in the design of the network. One is that the parameter learning at each hidden layer is given utilizing closed-form solutions, which is different from the BP algorithm with iterative updating strategy. Another aspect is that fast sparse deep neural networks use the summation method of a multi-layer linear approximation to estimate the output target, which is a different way from most deep neural network models. Unlike the traditional deep neural networks, fast sparse deep neural networks can achieve excellent generalization performance without fine-tuning. In addition, it is worth noting that fast sparse deep neural networks can also effectively overcome the shortcomings of the extreme learning machine and hierarchical extreme learning machine. Compared to the existing deep neural networks, enough experimental results on benchmark datasets demonstrate that the proposed model and optimization algorithms are feasible and efficient. INDEX TERMS Sparse representation, extreme learning machine, deep neural networks, convex approximation, fast sparse deep neural networks.

show abstract

Section: Introductionmentioning

confidence: 89%

Fast Sparse Deep Neural Networks: Theory and Performance Analysis

Zhao

Jiao

2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…We present the schematic illustration process of the work proposed iDRP framework, including the realization of the universal approximations of Multilayer Perceptron (MLP) with multiple hidden layers by exploiting the capabilities of H-ELM and other related techniques [30]- [33], which have the preeminent underexplored potentials for accelerated speed, rapid feature learning, and improved classification performance [30], [31], [34], [35]. Figure 1 Shows the schematic illustration of the setup of the experimental workflow process of the iDRP framework.…”

Section: Methodsmentioning

confidence: 99%

“…The Data Augmentation techniques and propagation techniques [30], [31], [37], which helped with the 'fine-grained extraction and classification' of signatures of complex datasets were applied to the model implementation of the proposed novel iDRP framework. The applied research approach for the development of the iDRP framework is implemented using Python Programming language (https://www.python.org/), which is an open-source language with rich and extensive libraries that is easily accessible to the general masses.…”

Section: Conceptual Implementation Of Proposed Frameworkmentioning

confidence: 99%

“…A formular for the H-ELM was derived from [30], [31] as a result of the amalgamation of the Radial Bias Function (RBF) nodes with the defined Activation function [30] for mapping indiscriminate features of the converted logged datasets of an IoT ecosystem. The primary aim is to precisely approximate the continuous targets with mobility at the evolving RBF nodes adaptively [30], [31], [37]. This ensures the irregular initialization of the hidden nodes H in a given set of big data in an IoT ecosystem.…”

Section: H-elm Derivative Formulamentioning

confidence: 99%

“…Evidently, the aforementioned complex IoT malware detection approaches are no longer suitable for current growing subtle and complex obfuscated cybersecurity paradigm for five (5) major reasons: (1) the distinct fragments (sections) of converted IoT malware image file can be visualized [27], [38]; (2) the distinct textural patterns of the converted IoT malware image files can be classified (and reclassified), and analyzed using image processing [27], [45] classification techniques commonly applied in computer vision [30], [31]; (3) the advanced attackers typically modify or reuse codes to generate new variants of complex obfuscated IoT malware scripts [38]- [40]; (4) the vector structures of converted IoT malwares image files of the same family are similar [27], which means vectorized representation (i.e., faster speed) of the IoT malware image files will be sufficient to be fed directly as input into the similar ML-DL algorithms [30], [31] built for image and pattern recognitions [27], [45]- [47]; (5) the captured IoT malware dataset once converted to image files will be immutable (nondestructive and tamper proof), so there is no workaround its state by the attackers. Therefore, an amalgamation and syntheses of the modern ML-DL techniques and methods designed utilizing converted immutable IoT malware binaries to image files is both a logical and pragmatic solution to the current rapidly growing complex IoT malware problems that would scale adaptively.…”

Section: The Need To Convert Iot Malware Binaries To Imagesmentioning

confidence: 99%

See 2 more Smart Citations

iDRP Framework: An Intelligent Malware Exploration Framework for Big Data and Internet of Things (IoT) Ecosystem

Eboya¹,

Juremi²

2021

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

The Internet of Things (IoT) is at a face paced growth in the advanced Industrial Revolution (IR) 4.0 in the modern digital world. Considering the current network security challenges and sophistication of attacks in the heavily computerized and interconnected systems, such as an IoT ecosystem, the need for an innovative, robust, intelligent and adaptive malware attacks and threats security solution is becoming predominant in the current cyberspace. An integrated and scalable IoT malware detection framework called iDRP framework with deep learning method was proposed as a solution to current IoT malware attacks that are largely obfuscated. The novel framework utilized systematic pre-processing and post-processing techniques and methods on the BoTNetIoT malware datasets that contains both benign and malicious IoT traffic data infected by modern day IoT attacks such as Mirai and Gafgyt etc. IoT malware variants in an IoT ecosystem. The raw IoT malware binaries were converted to image files (Gray-scaled) and computed statistically with synthesised sparsed and differential evolutionary hidden feature structures techniques, which were cyclically trained, tested, and cross-validated to establish empirical anomalies with precision in the detection, recognizing, and prediction of malware anomalies in a modern IoT ecosystem. Preliminary experiments were conducted with standardized image binary files such as the datasets as sound scientific exploratory experiments with profound results. The comparative results of the performance of our integrated techniques and methods on the BoTNetIoT IoT malware datasets achieved a 99.98% accuracy, 99.99% ROC/AUC, 99.95% precision, and 99.93 recall rate etc. utilizing the integrated iDRP framework mechanisms for effectively detecting IoT malware in an IoT ecosystem.

show abstract