Detection and Mitigation of Rare Subclasses in Deep Neural Network Classifiers

Classification of Bisyllabic Lexical Stress Patterns Using Deep Neural Networks

Deep neural networks for texture classification—A theoretical analysis

We investigate the use of Deep Neural Networks for the classification of image datasets where texture features are important for generating class-conditional discriminative representations. To this end, we first derive the size of the feature space for some standard textural features extracted from the input dataset and then use the theory of Vapnik-Chervonenkis dimension to show that hand-crafted feature extraction creates low-dimensional representations which help in reducing the overall excess error rate. As a corollary to this analysis, we derive for the first time upper bounds on the VC dimension of Convolutional Neural Network as well as Dropout and Dropconnect networks and the relation between excess error rate of Dropout and Dropconnect networks. The concept of intrinsic dimension is used to validate the intuition that texture-based datasets are inherently higher dimensional as compared to handwritten digits or other object recognition datasets and hence more difficult to be shattered by neural networks. We then derive the mean distance from the centroid to the nearest and farthest sampling points in an n-dimensional manifold and show that the Relative Contrast of the sample data vanishes as dimensionality of the underlying vector space tends to infinity.

Heartbeat classification is a crucial step for arrhythmia diagnosis during electrocardiographic (ECG) analysis. The new scenario of wireless body sensor network- (WBSN-) enabled ECG monitoring puts forward a higher-level demand for this traditional ECG analysis task. Previously reported methods mainly addressed this requirement with the applications of a shallow structured classifier and expert-designed features. In this study, modified frequency slice wavelet transform (MFSWT) was firstly employed to produce the time-frequency image for heartbeat signal. Then the deep learning (DL) method was performed for the heartbeat classification. Here, we proposed a novel model incorporating automatic feature abstraction and a deep neural network (DNN) classifier. Features were automatically abstracted by the stacked denoising auto-encoder (SDA) from the transferred time-frequency image. DNN classifier was constructed by an encoder layer of SDA and a softmax layer. In addition, a deterministic patient-specific heartbeat classifier was achieved by fine-tuning on heartbeat samples, which included a small subset of individual samples. The performance of the proposed model was evaluated on the MIT-BIH arrhythmia database. Results showed that an overall accuracy of 97.5% was achieved using the proposed model, confirming that the proposed DNN model is a powerful tool for heartbeat pattern recognition.

In the past decade, Deep Neural Networks (DNNs) have achieved breakthrough collaborations in developing smart intelligent systems within the field of computer vision, natural language processing, autonomous systems, etc. Recent research has revealed that stability of such smart systems is at greater risk when they come across to adversarial perturbations. Although, these perturbations may not be perceivable in nature when seen from naked eye, yet, they are capable enough to fool state-of-the-art DNN classifiers. Till now, much of the previous work related to fool such classifiers focuses on generating adversaries that directly change pixel values of an image in spatial-domain. In this paper, we propose a novel transform-domain imperceptible attack methodology “TDIAM” to generate adversaries based on image steganography-approach using a “single carefully selected targeted watermark”. We use three different frequency-domain approaches, i.e., Discrete Wavelet Transform (DWT), Discrete Cosine Transform (DCT) and Fast Fourier Transform (FFT) to craft perturbations in selective frequency component which makes it robust and it requires less computational time as it is a non-gradient approach. We present our case study on MNIST handwritten digits dataset. Our results demonstrate that the generated perturbation vector successfully fool simple Convolutional Neural Network (CNN), LeNet-5 and AlexNet architectures by increasing probability of adversarial examples for the targeted class (to which the targeted watermark belongs) in both “black-box” and “white-box” adversarial attacks. The results have shown that among these three perturbation approaches, DWT based perturbation shown promising results by effectively fooling DNNs while ensuring the high imperceptibility as well.

Learning valid similarities is a vital problem in hashing methods, especially in large-scale image search. Similarity pertained hashing method is widely used in image retrieval for its high quality compact binary code mapping. The hashing scheme of most existing hashing methods is that the input data is encoded as a vector of visual features and hashed into binary hash codes via projection functions or quantization methods afterward. However, this separated pipeline may prone to lose accurate similarities of images, since the limited compatible domain between visual feature vectors generation and binary codes mapping process. Encouraged by the extraordinary image representation learning ability of deep neural networks in classification, we propose a structure that merges binary code generation process within deep neural networks for efficient image retrieval. The proposed architecture contains two fundamental blocks. The stacked convolution layers of Network In Network with global average pooling compute effective image representation and the embedded latent layer with binary activation functions learn binary hash codes simultaneously. Experiments show that the proposed method gains improvement over several state-of-the-art hashing methods.

Three-terminal vertical ferroelectric synaptic barristor enabled by HZO/graphene heterostructure with rebound depolarization

Due to the advent of modern embedded systems and mobile devices with constrained resources, there is a great demand for incredibly efficient deep neural networks for machine learning purposes. There is also a growing concern of privacy and confidentiality of user data within the general public when their data is processed and stored in an external server which has further fueled the need for developing such efficient neural networks for real-time inference on local embedded systems. The scope of our work presented in this paper is limited to image classification using a convolutional neural network. A Convolutional Neural Network (CNN) is a class of Deep Neural Network (DNN) widely used in the analysis of visual images captured by an image sensor, designed to extract information and convert it into meaningful representations for real-time inference of the input data. In this paper, we propose a neoteric variant of deep convolutional neural network architecture to ameliorate the performance of existing CNN architectures for real-time inference on embedded systems. We show that this architecture, dubbed CondenseNeXt, is remarkably efficient in comparison to the baseline neural network architecture, CondenseNet, by reducing trainable parameters and FLOPs required to train the network whilst maintaining a balance between the trained model size of less than 3.0 MB and accuracy trade-off resulting in an unprecedented computational efficiency.

Detecting pneumonia is a demanding task which always requires looking at chest X-ray images of patients suffering from it. The normal chest X-ray, reveals clear lungs without any areas of abnormal obscurity or opacification in the image. The Bacterial pneumonia, regularly consists of a focal lobar consolidation (a lung tissue which is filled with liquid instead of air) whereas the viral pneumonia shows distinct and more diffuse ‘interstitial’ pattern in both lungs. These characteristics can be identified, only with the help of an experienced radiologist. But these characteristics of pneumonia can also be overlapped with other diseases, further complicating the diagnosis. Our objective is to build a Convolutional Neural Network (CNN, or ConvNet) classifier to detect pneumonia in X-ray images of the patient. The dataset for the images is taken from kaggle—a data science learning and competition platform. Convolutional Neural Network (CNN or ConvNet) is a class of deep neural networks that specialises in analysing images and thus is widely used in computer vision applications such as image classification and clustering, object detection and neural style transfer. The above-stated objective will be implemented using Python as a programming language and using concepts of deep learning and neural networks. The dataset consists of 3 folders (train, test, val) and contains subfolders for each image category (Pneumonia/Normal). There are approximately 5000 X-Ray images (JPEG) and 2 categories (Pneumonia/Normal).

SummaryLung cancer is one of the major causes of death in the world, according to radiologists. However, a constant flow of medical images to hospitals is forcing radiologists to focus on accurate early prediction of nodules. Recently, several image‐processing techniques have cooperated for the early prediction of lung nodules. However, it's hard to detect strong nodes because of lung node diversity and environmental complexity. This study presents a hybrid machine learning technique for predicting an early prognosis of lung nodules from clinical images using a learning‐based neural network classifier. First, we introduce an improved snake swarm optimization with a bat model (ISSO‐B) for lung nodule segmentation using statistical information. Second, we demonstrate a chaotic atom search optimization (CASO) algorithm to select the optimal best features among multiple features, which minimize the dimensionality problem. Third, we develop a hybrid learning‐based deep neural network classifier (L‐DNN) for nodule prediction and classification. Finally, we evaluate our proposed technique with different public datasets LIDC‐IDRI and FAH‐GMU. Then, performance can be compared with the latest technology in terms of accuracy, sensitivity, specificity, and area under curve (AUC).

The continuous rise of the COVID-19 Omicron cases despite the vaccination program available has been progressing worldwide. To mitigate the COVID-19 contraction, different contact tracing applications have been utilized such as Thai Chana from Thailand. This study aimed to predict factors affecting the perceived usability of Thai Chana by integrating the Protection Motivation Theory and Technology Acceptance Theory considering the System Usability Scale, utilizing deep learning neural network and random forest classifier. A total of 800 respondents were collected through convenience sampling to measure different factors such as understanding COVID-19, perceived severity, perceived vulnerability, perceived ease of use, perceived usefulness, attitude towards using, intention to use, actual system use, and perceived usability. In total, 97.32% of the deep learning neural network showed that understanding COVID-19 presented the most significant factor affecting perceived usability. In addition, random forest classifier produced a 92% accuracy with a 0.00 standard deviation indicating that understanding COVID-19 and perceived vulnerability led to a very high perceived usability while perceived severity and perceived ease of use also led to a high perceived usability. The findings of this study could be considered by the government to promote the usage of contact tracing applications even in other countries. Finally, deep learning neural network and random forest classifier as machine learning algorithms may be utilized for predicting factors affecting human behavior in technology or system acceptance worldwide.

In pattern recognition, identifying Kannada handwritten numerals are a complex knot. This paper portrays an avenue that pikes us to attain a most potent Kannada Numerals recognition process. In this, handwritten Kannada characters are captivated in document fashion and are subjected to Pre-processing and attribute extraction processes. Pre-processing entails steps like noise removal, binarization, normalization, skew amendment, and thinning. Features are extricated by exploiting strategies like Drift Length Count, Direction related progression code, DWT and Curvelet Transfiguration Wrapping. For an impressive classification process deep convolution neural network classifier is preferred. Isolation accuracy of Kannada numeral aimed here will outsource 96% of accuracy.

Neutrinos, dark matter, and long-lived neutral particles traverse the particle detectors unnoticed, carrying away information about their parent particles and interaction sources needed to reconstruct key variables like resonance peaks in invariant mass distributions. In this work, we show that a k-nearest neighbors regressor algorithm combined with deep neural network classifiers, a kNNNN, is able to accurately recover binned distributions of the fully leptonic WW mass of a new heavy Higgs boson and its Standard Model backgrounds from the observable detector level information at disposal. The output of the regressor can be used to train even stronger classifiers to separate signals and backgrounds in the fully leptonic case and guarantee the selection of on-mass-shell Higgs bosons with enhanced statistical significance. The method assumes previous knowledge of the event classes and model parameters, thus suitable for post-discovery studies.

This paper presents a distributed modeling approach of deep neural network (DNN) classifier for segmentation-free Telugu word recognition. The DNN is a hybrid neural network model. It involves the convolutional neural network (CNN) and recurrent neural network (RNN) layers. Moreover, the connectionist temporal classification (CTC) layer gives the output. The presence of CNN layers requires advanced computer systems consisting of graphics processing units (GPUs) for speedy training of a DNN model. The distributed modeling approach involves the distribution of load across a TPU. As a comparative study, we also train the same model on CPU and GPU systems. Our method aims to handle large batch sizes and reduce modeling time, which is precious to researchers. The word error rate (WER) of the test set is 5%, which is promising.KeywordsCNNCTCDistributed ModelingDNNGPULSTMRNNTelugu OCRWord Recognition

The rise of cultural tourism in India and massive digitization over the last decade has necessitated preserving Indian art forms. Recent advances in artificial intelligence (AI) have provided the tools and techniques to make these efforts more efficient. The potential research areas involved are digital image processing, computer vision, and machine intelligence. Since researchers have not much explored the scope of the digital preservation of Indian classical dance, this paper contributes a method to classify different Indian classical dance forms. Computerized dance modeling offers a wide range of uses, including cultural conservation, online learning, entertainment, and even remedial treatment. Because the dataset is small, the paper uses transfer learning models for feature extraction and deep neural networks for classification. The proposed method showed a remarkable accuracy of 97.05 percent.