Search

In this paper, the classification of Unmanned Aerial Vehicles (UAV) data traffic into three distinct classes: analog video, digital OFDM-modulated video, and Additive White Gaus-sian Noise (AWGN) has been performed employing six neural network classifiers including Feed Forward Neural Network (FFNN), Generalized Regression Neural Network (GRNN), and Probabilistic Neural Network (PNN); and Cascade Forward Neural Network (CFNN), Recurrent Neural Network (RNN) and multilayer perceptron neural network (NN). The data set composed of the time domain signal samples for classifiers' training, validation, and testing has been collected in the controlled exper-iment conducted in the office/lab environment with the stationary signal source and receiver. The subset of twenty-four extracted features has been used as input to the neural network classifiers. Feature reduction has been performed using four popular in literature feature selection algorithms: Minimum Redundancy Maximum Relevance (MRMR), Neighborhood Component Anal-ysis (NCA), Relief, and Laplacian score to enhance computational efficiency and prediction speed for hardware implementation and real-time operation on the target CPU. Four features including mean, standard deviation, and median absolute deviation of the time domain signal, and RSSI have been selected. Six neural network classifiers have been trained using both the full and reduced feature sets. Also, two validation algorithms: k-fold cross-validation and hold-out validation have been evaluated. The Recurrent Neural Network (RNN) has demonstrated the highest accuracy using the full feature set and employing cross-validation. The feature reduction has led to a 3 % decrease in accuracy for RNN. Feedforward Neural Network (FFNN) has demonstrated the highest accuracy of 93.51 % with the reduced feature set input using cross-validation on PC in Matlab environment. It has been prototyped on our target hardware CPU using Mathworks Embedded Coder; the generated C c

16th International Conference on Electronics Computer and Computation, ICECCO 2021 -- 25 November 2021 through 26 November 2021 -- -- 176250, Application of deep learning techniques for breast cancer classification using histopathology images has gained interest during recent years. In this study, an open-source convolutional neural network (CNN) model developed for breast cancer classification model is optimized by performing sensitivities on various CNN parameters such as data balancing, activation functions and adding/removing CNN layers. Some of the parameters are less-sensitive in affecting model's performance. The results show that by balancing the number of positive and negative samples, accuracy of the model can be improved. However, some additional work is required to reach to that point. Furthermore, the computation time is reduced by almost 30% by increasing the learning rate from 0.01 to 0.05 while the training and validation accuracy and loss are comparable to that of the original CNN model. © 2021 IEEE., Nazarbayev University, NU, V. ACKNOWLEDGEMENT This work was supported by Faculty Development Competitive Research Grant Program (No. 080420FD1909) at Nazarbayev University.

29th IEEE Conference on Signal Processing and Communications Applications, SIU 2021 -- 9 June 2021 through 11 June 2021 -- 170536, In this study, our primary aim is to detect different formations, objects on the images taken from various underwater videos. For this purpose, machine learning models such as SVM, multi-layer perceptron, logistic regression that use attributes, image histogram obtained from images were chosen. In addition, Autoencoder and CNN based deep learning models were used directly over images and their performances were compared. According to the results, it was observed that all models were satisfactory and achieved good classification performances. The highest performance was observed in the Autoencoder based deep learning model, which achieved an accuracy level of %95. In the future, we are planning to continue studies to focus on underwater cable tracking and detecting errors and anomalies in underwater cables. © 2021 IEEE.

16th International Conference on Electronics Computer and Computation, ICECCO 2021 -- 25 November 2021 through 26 November 2021 -- -- 176250, Application of deep learning techniques for breast cancer classification using histopathology images has gained interest during recent years. In this study, an open-source convolutional neural network (CNN) model developed for breast cancer classification model is optimized by performing sensitivities on various CNN parameters such as data balancing, activation functions and adding/removing CNN layers. Some of the parameters are less-sensitive in affecting model's performance. The results show that by balancing the number of positive and negative samples, accuracy of the model can be improved. However, some additional work is required to reach to that point. Furthermore, the computation time is reduced by almost 30% by increasing the learning rate from 0.01 to 0.05 while the training and validation accuracy and loss are comparable to that of the original CNN model. © 2021 IEEE., Nazarbayev University, NU, V. ACKNOWLEDGEMENT This work was supported by Faculty Development Competitive Research Grant Program (No. 080420FD1909) at Nazarbayev University.

Recently, there is increasing interest in detecting electricity thieves for economic benefits for power companies, and many works aim to improve the accuracy of electricity theft detection. Nevertheless, a core obstacle that currently hinders the direct comparison of classifiers for electricity theft detection is the lack of a standard and public dataset, since fraudulent power load profiles are usually difficult to collect for various reasons, including cost, cumber, and confidentiality. Therefore, this paper presents an open-source toolbox, which generates different kinds of fraudulent power load profiles from attack models, and integrates classical classifiers (e.g., support vector machine, multi-layer perceptron, convolutional neural network, long short-term memory, bidirectional long short-term memory) with different performance as baselines for the comparison with new algorithms. Users can easily generate datasets and modify parameters of classical classifiers guided by user friendly interactive interfaces. The codes, toolbox, and user manuals are available online and it is free to use and extend them., Part of proceedings: ISBN 978-1-6654-0064-0QC 20220816

This work presents InfiniWolf, a novel multi-sensor smartwatch that can achieve self-sustainability exploiting thermal and solar energy harvesting, performing computationally high demanding tasks. The smartwatch embeds both a System-on-Chip (SoC) with an ARM Cortex-M processor and Bluetooth Low Energy (BLE) and Mr. Wolf, an open-hardware RISC-V based parallel ultra-low-power processor that boosts the processing capabilities on board by more than one order of magnitude, while also increasing energy efficiency. We demonstrate its functionality based on a sample application scenario performing stress detection with multi-layer artificial neural networks on a wearable multi-sensor bracelet. Experimental results show the benefits in terms of energy efficiency and latency of Mr. Wolf over an ARM Cortex-M4F micro-controllers and the possibility, under specific assumptions, to be self-sustainable using thermal and solar energy harvesting while performing up to 24 stress classifications per minute in indoor conditions. © 2020 EDAA.

Multilayer Perceptron (MLP) is used in a broad range of applications. Hardware acceleration of MLP is one most promising way to provide better performance-energy efficiency. Previous works focused on the intra-layer optimization and layer-after-layer processing, while leaving the inter-layer optimization never studied. In this paper, we propose hardware acceleration of MLPs based on inter-layer optimization which allows us to overlap the execution of MLP layers. First we describe the inter-layer optimization from software and mathematical perspectives. Then, a reference Two-Neuron architecture which is efficient to support the inter-layer optimization is proposed and implemented. Discussions about area cost, performance and energy consumption are carried out to explore the scalability of the Two-Neuron architecture. Results show that the proposed MLP design optimized across layers achieves better performance and energy efficiency than the conventional intra-layer optimized designs. As such, the inter-layer optimization provides another possible direction other than the intra-layer optimization to gain further performance and energy improvements for the hardware acceleration of MLPs., QC 20200626

Deep neural networks show high accuracy in the problem of semantic and instance segmentation of biomedical data. However, this approach is computationally expensive. The computational cost may be reduced with network simplification after training or choosing the proper architecture, which provides segmentation with less accuracy but does it much faster. In the present study, we analyzed the accuracy and performance of UNet and ENet architectures for the problem of semantic image segmentation. In addition, we investigated the ENet architecture by replacing of some convolution layers with box-convolution layers. The analysis performed on the original dataset consisted of histology slices with mast cells. These cells provide a region for segmentation with different types of borders, which vary from clearly visible to ragged. ENet was less accurate than UNet by only about 1-2%, but ENet performance was 8-15 times faster than UNet one. © 2019 IEEE.