Double-Shift: A Low-Power DNN Weights Storage and Access Framework based on Approximate Decomposition and Quantization

One major challenge in deploying Deep Neural Network (DNN) in resource-constrained applications, such as edge nodes, mobile embedded systems, and IoT devices, is its high energy cost. The emerging approximate computing methodology can effectively reduce the energy consumption during the computing process in DNN. However, a recent study shows that the weight storage and access operations can dominate DNN's energy consumption due to the fact that the huge size of DNN weights must be stored in the high-energy-cost DRAM. In this paper, we propose Double-Shift, a low-power DNN weight storage and access framework, to solve this problem. Enabled by approximate decomposition and quantization, Double-Shift can reduce the data size of the weights effectively. By designing a novel weight storage allocation strategy, Double-Shift can boost the energy efficiency by trading the energy consuming weight storage and access operations for low-energy-cost computations. Our experimental results show that Double-Shift can reduce DNN weights to 3.96%–6.38% of the original size and achieve an energy saving of 86.47%–93.62%, while introducing a DNN classification error within 2%.

Age group classification and gender recognition from speech with temporal convolutional neural networks

This paper analyses the performance of different types of Deep Neural Networks to jointly estimate age and identify gender from speech, to be applied in Interactive Voice Response systems available in call centres. Deep Neural Networks are used, because they have recently demonstrated discriminative and representation capabilities in a wide range of applications, including speech processing problems based on feature extraction and selection. Networks with different sizes are analysed to obtain information on how performance depends on the network architecture and the number of free parameters. The speech corpus used for the experiments is Mozilla’s Common Voice dataset, an open and crowdsourced speech corpus. The results are really good for gender classification, independently of the type of neural network, but improve with the network size. Regarding the classification by age groups, the combination of convolutional neural networks and temporal neural networks seems to be the best option among the analysed, and again, the larger the size of the network, the better the results. The results are promising for use in IVR systems, with the best systems achieving a gender identification error of less than 2% and a classification error by age group of less than 20%.

DATA MINING ALGORITHM C4.5 CLASSIFICATION DETERMINATION CREDIT ELIGIBILITY FOR JAYA BERSAMA COOPERATIVES (KORJABE)

This study uses the C4.5 classification algorithm to determine creditworthness, clasification aims to divide the assigned object intoin a number of categories called classes. In this study, the authorusing data mining and C4.5 algorithm as the selection method. The criteria used are loan installments, prospective customer income, termloan time, status of prospective customers. This study resulted in a classification modeldecision tree using the C4.5 algorithm is included in the Excellent category Classification with an accuracy value of 98.33% and a classification error of 1.67%,so that this study uses 70% training data and 30% test data. From resultthe calculation obtained shows that the C4.5 algorithm can be usedto determine the feasibility of granting credit to Koperasi Jaya customers Together (KORJABE). Keywords: Analysis, Credit Eligibility, C4 Algorithm, Data Mining, Method Abstrak: Penelitian ini menggunakan metode Algoritma C4.5 klasifikasi untuk menentukan kelayakan kredit, klasifikasi bertujuan untuk membagi objek yang ditetapkan ke dalam satu nomor kategori yang disebut kelas. Dalam penelitian ini, penulis menggunankan data mining dan algoritma C4.5 sebagai metode pemilihannya. Kriteria yang digunakan yaitu , angsuran pinjaman,penghasilan calon nasabah,jangka waktu pinjaman ,status calon nasabah. Penelitian ini menghasillkan model klasifikasi pohon keputusan menggunakan algoritma C4.5 termasuk dalam kategori Excellent Classification dengan nilai akurasi sebesar 98,33% dan klasifikasi eror 1,67%, sehingga penelitian ini kan menggunakan data latih 70% dan data uji 30%. Dari hasil perhitungan yang diperoleh menunjukan bahwa algoritma C4.5 dapat digunakan untuk menen tukan kelayakan pemberian kredit kepada nasabah Koperasi Jaya Bersama (KORJABE). Kata kunci: Algoritma C4.5, Analisis, Data Mining, Kelayakan Kredit, Metode

Clustering by Errors: A Self-Organized Multitask Learning Method for Acoustic Scene Classification

Acoustic scene classification (ASC) tries to inference information about the environment using audio segments. The inter-class similarity is a significant issue in ASC as acoustic scenes with different labels may sound quite similar. In this paper, the similarity relations amongst scenes are correlated with the classification error. A class hierarchy construction method by using classification error is then proposed and integrated into a multitask learning framework. The experiments have shown that the proposed multitask learning method improves the performance of ASC. On the TUT Acoustic Scene 2017 dataset, we obtain the ensemble fine-grained accuracy of 81.4%, which is better than the state-of-the-art. By using multitask learning, the basic Convolutional Neural Network (CNN) model can be improved by about 2.0 to 3.5 percent according to different spectrograms. The coarse category accuracies (for two to six super-classes) range from 77.0% to 96.2% by single models. On the revised version of the LITIS Rouen dataset, we achieve the ensemble fine-grained accuracy of 83.9%. The multitask learning models obtain an improvement of 1.6% to 1.8% compared to their basic models. The coarse category accuracies range from 94.9% to 97.9% for two to six super-classes with single models.

A Dual-Stage Vocabulary of Features (VoF)-Based Technique for COVID-19 Variants’ Classification

Novel coronavirus, known as COVID-19, is a very dangerous virus. Initially detected in China, it has since spread all over the world causing many deaths. There are several variants of COVID-19, which have been categorized into two major groups. These groups are variants of concern and variants of interest. Variants of concern are more dangerous, and there is a need to develop a system that can detect and classify COVID-19 and its variants without touching an infected person. In this paper, we propose a dual-stage-based deep learning framework to detect and classify COVID-19 and its variants. CT scans and chest X-ray images are used. Initially, the detection is done through a convolutional neural network, and then spatial features are extracted with deep convolutional models, while handcrafted features are extracted from several handcrafted descriptors. Both spatial and handcrafted features are combined to make a feature vector. This feature vector is called the vocabulary of features (VoF), as it contains spatial and handcrafted features. This feature vector is fed as an input to the classifier to classify different variants. The proposed model is evaluated based on accuracy, F1-score, specificity, sensitivity, specificity, Cohen’s kappa, and classification error. The experimental results show that the proposed method outperforms all the existing state-of-the-art methods.

Feature Optimization of Exhaled Breath Signals Based on Pearson-BPSO

Feature optimization, which is the theme of this paper, is actually the selective selection of the variables on the input side at the time of making a predictive kind of model. However, an improved feature optimization algorithm for breath signal based on the Pearson-BPSO was proposed and applied to distinguish hepatocellular carcinoma by electronic nose (eNose) in the paper. First, the multidimensional features of the breath curves of hepatocellular carcinoma patients and healthy controls in the training samples were extracted; then, the features with less relevance to the classification were removed according to the Pearson correlation coefficient; next, the fitness function was constructed based on K-Nearest Neighbor (KNN) classification error and feature dimension, and the feature optimization transformation matrix was obtained based on BPSO. Furthermore, the transformation matrix was applied to optimize the test sample’s features. Finally, the performance of the optimization algorithm was evaluated by the classifier. The experiment results have shown that the Pearson-BPSO algorithm could effectively improve the classification performance compared with BPSO and PCA optimization methods. The accuracy of SVM and RF classifier was 86.03% and 90%, respectively, and the sensitivity and specificity were about 90% and 80%. Consequently, the application of Pearson-BPSO feature optimization algorithm will help improve the accuracy of hepatocellular carcinoma detection by eNose and promote the clinical application of intelligent detection.

An Efficient Machine Learning based Model for Classification of Wearable Clothing

Computer vision research and its applications in the fashion industry have grown popular due to the rapid growth of information technology. Fashion detection is increasingly popular because most fashion goods need detection before they could be worn. Early detection of the human body component of the input picture is necessary to determine where the garment area is and then synthesize it. For this reason, detection is the starting point for most of the in-depth research. The cloth detection of landmarks is retrieved through many feature items that emphasis on fashionate things. The feature extraction can be done for better accuracy, pose and scale transmission. These convolution filters extract the features through many epochs and max-pooling layers in the neural networks. The optimized classification has been done using SVM in this study, for attaining overall high efficiency. This proposed CNN approach fashionate things prediction is combined with SVM for better classification. Furthermore, the classification error is minimized through the evaluation procedure for obtaining better accuracy. Finally, this research work has attained good accuracy and other performance metrics than the different traditional approaches. The benchmark datasets, current methodologies, and performance comparisons are all reorganized for each piece.

Pigeon M etheuristic Optimized Generative Adversarial Networks and ARKFCM Algorithms for retinal V essel Segmentation and Classification

Automatic evaluation of retinal vessels acts a significant part in diagnosis of several ocular and systemic diseases. Eye diseases must be diagnosed early to avoid severe infection and vision loss. The method of segmentation and classification of the retinal blood vessel identification is most difficult tasks in computerized fundus imaging now a days. To solve this problem in this paper, to locate retinal vessel in the retinal vessel, Adaptive Regularized Kernel Based Fuzzy Clustering Means (ARKFCM) algorithm-based segmentation is used. For retinal vessel prediction purpose in this paper a PIGEON optimization-based learning rate modified Generative Adversarial Networks (GAN) algorithm is introduced. Additionally, to improve the proposed classification performance input image is transformed with the aid of Discrete Wavelet Transform (DWT). The DWT applied Low Low (LL) image and segmented images are cascaded. The cascade images are used for training and testing. The proposed system has validated with the help of DRIVE and STARE publically available datasets. They are studied by applying a Convolutional Neural Network, an instantly trained neural network for predicting retinal vessel. In the end, the system is checked for system efficiency using the results of modeling based on MATLAB. The scheme guarantees an accuracy of 92.77% on DRIVE dataset and 98.85% on STARE dataset with a minimum average classification error of 2.57%. Further, we recommended to physician for implement the real time clinical application; this scheme is highly beneficial for doctors for identifying retinal blood vessels.

Error rate control for classification rules in multiclass mixture models

In the context of finite mixture models one considers the problem of classifying as many observations as possible in the classes of interest while controlling the classification error rate in these same classes. Similar to what is done in the framework of statistical test theory, different type I and type II-like classification error rates can be defined, along with their associated optimal rules, where optimality is defined as minimizing type II error rate while controlling type I error rate at some nominal level. It is first shown that finding an optimal classification rule boils down to searching an optimal region in the observation space where to apply the classical Maximum A Posteriori (MAP) rule. Depending on the misclassification rate to be controlled, the shape of the optimal region is provided, along with a heuristic to compute the optimal classification rule in practice. In particular, a multiclass FDR-like optimal rule is defined and compared to the thresholded MAP rules that is used in most applications. It is shown on both simulated and real datasets that the FDR-like optimal rule may be significantly less conservative than the thresholded MAP rule.

Mental Stress Classification Based on a Support Vector Machine and Naive Bayes Using Electrocardiogram Signals

Examining mental health is crucial for preventing mental illnesses such as depression. This study presents a method for classifying electrocardiogram (ECG) data into four emotional states according to the stress levels using one-against-all and naive Bayes algorithms of a support vector machine. The stress classification criteria were determined by calculating the average values of the R-S peak, R-R interval, and Q-T interval of the ECG data to improve the stress classification accuracy. For the performance evaluation of the stress classification model, confusion matrix, receiver operating characteristic (ROC) curve, and minimum classification error were used. The average accuracy of the stress classification was 97.6%. The proposed model improved the accuracy by 8.7% compared to the previous stress classification algorithm. Quantifying the stress signals experienced by people can facilitate a more effective management of their mental state.