Representation Of Input Data Research Articles

Recurrent Neural Networks based Online Behavioural Malware Detection Techniques for Cloud Infrastructure

: With the increased basis of applications and services on Cloud, the dependence on cloud technologies has expanded rapidly. This transition eases some of the hardware management troubles, but it also exposes broad security threats, especially in regard to malware. Malware is a major threat to cloud computing systems mainly in IaaS dominated environments. In this paper we explore the usage of Recurrent Neural Networks, RNNs in malware detection in cloud VMs. In particular, we examine two of the most popular RNN models; Long short term Memory, LSTM and Bidirectional RNNs, BIDI, that try to understand malware patterns by watching the activity of systems and the use of CPU, Memory and Disk. On the dataset of 40,680 samples 40,680 malware and benign samples, the models were trained on process level features from real malware in a free and open cloud environment. Both models achieved detection rates of over 99% in average key performance metrics. Moreover, wealso investigate the impact of different representations of input data on the performance of the models. The results of the study clearly support the viability of RNNs for online malware detection in cloud infrastructure, and allow for high accuracy and real time detection.

Development of a Method for Optimization of the Process of Extraction of Values of Numerical Characteristics of Oil Feedstock Components from Graphical Data of Chromatographic Analysis

A variant of the algorithm has been developed to perform the procedure of automated recovery of numerical values of graphically represented chromatograph signal function, studying the component composition of heavy oil feedstock samples. The problem, which the developed method aims to solve, consists in the poor adaptation of chromatographs to the oil industry: oil is a natural raw material, which is not chemically pure, therefore not all numerical characteristics of the components contained in the investigated sample are fixed within the chromatographic study. In the current configuration of the method, the values from the chromatogram are recorded manually. The developed method takes as input data the images of oil chromatograms obtained in the laboratory, presented in the original black and white colour scheme. The output data of the method is an array of numerical values of coordinates reconstructed with a step of one pixel. The size of the error in the reconstruction of the values by the method is much smaller than the threshold set by the petrochemical laboratory. In addition to automating the indicated task, the array of obtained coordinate values was vectorized in order to use the vector as input data in the Transformer model to solve the problem of predicting the redistribution of hydrocarbon components of heavy oil under the influence of catalysts. As a result of the change in input data representation, the time required to obtain a prediction and the training time were reduced by a multiple, while the value of the average prediction error decreased.

Multimodal Emotion Recognition Using Visual, Vocal and Physiological Signals: A Review

The dynamic expressions of emotion convey both the emotional and functional states of an individual’s interactions. Recognizing the emotional states helps us understand human feelings and thoughts. Systems and frameworks designed to recognize human emotional states automatically can use various affective signals as inputs, such as visual, vocal and physiological signals. However, emotion recognition via a single modality can be affected by various sources of noise that are specific to that modality and the fact that different emotion states may be indistinguishable. This review examines the current state of multimodal emotion recognition methods that integrate visual, vocal or physiological modalities for practical emotion computing. Recent empirical evidence on deep learning methods used for fine-grained recognition is reviewed, with discussions on the robustness issues of such methods. This review elaborates on the profound learning challenges and solutions required for a high-quality emotion recognition system, emphasizing the benefits of dynamic expression analysis, which aids in detecting subtle micro-expressions, and the importance of multimodal fusion for improving emotion recognition accuracy. The literature was comprehensively searched via databases with records covering the topic of affective computing, followed by rigorous screening and selection of relevant studies. The results show that the effectiveness of current multimodal emotion recognition methods is affected by the limited availability of training data, insufficient context awareness, and challenges posed by real-world cases of noisy or missing modalities. The findings suggest that improving emotion recognition requires better representation of input data, refined feature extraction, and optimized aggregation of modalities within a multimodal framework, along with incorporating state-of-the-art methods for recognizing dynamic expressions.

All signals point to personality: A dual-pipeline LSTM-attention and symbolic dynamics framework for predicting personality traits from Bio-Electrical signals

The prediction of personality traits offers valuable insights into human behaviour, more specifically in psychology, healthcare, and social science. In this paper, we present a novel methodology for personality trait prediction using a dual-pipeline architecture. The model architecture leverages Long Short-Term Memory (LSTM) networks with batch normalization for capturing sequential dependencies in data and incorporates temporal attention heads for feature extraction. By combining these parallel pipelines, our network effectively utilizes both LSTM and attention mechanisms to create a comprehensive representation of input data. The network’s goal is to predict the OCEAN (openness, conscientiousness, extraversion, agreeableness and neuroticism) traits using physiological signals including: EEG, ECG and GSR. Including attention mechanisms enables the model to focus on critical moments in these signals, resulting in significantly improved prediction accuracy. Experimental evaluations demonstrate the superior performance of our method compared to traditional machine learning methods on two publicly available datasets: ASCERTAIN and AMIGOS. Our source code is accessible at https://github.com/deepakkumar-iitr/AT3NET.

Применение нейроморфных электроактивных макроячеек для повышения надежности пространственной ориентации

Neural network architecture based on polymer macrocells acting as neural elements and capable to generate electrically conductive synaptic junctions is developed. The architecture of a neural network processor where each component can record and produce a response to energy effects of different nature: electrical, optical and electrostatic has been developed. This function of the neural network eliminates the limitations related to the perception of the input information only by the input layer of the neural network processor. The impact of the influence of different input data representations on the information processing by the macrocell has been studied. The efficiency of temporal and electrical degradation of a neural network macrocell as a complex phenomenon directed on adaptation of a neural network to this process and generation of an algorithm for distribution of elements of a tutorial sample evenly over the total volume of available neural cells has been studied. The article studies a method of generation of episodic circulating memory, which significantly improves the speed of generation of a solution by the network, taking into account the gravitational interaction with macrocells. The revealed specific nature of the influence of the gravitational field on the functioning of neural network clusters was the basis for natural selection of cells by the performed properties depending on the spatial position in the neural network architecture. A method of exposure of the cell to physical equivalents of tutorial sample elements converted into different energy patterns to control the configuration of the macrocell using all available modes of information perception has been provided. Each exposure containing information on the tutorial element (optical and electrostatic) supports the generated generalising abilities of the neural macrocells composition. Practical significance of the performed research comprises the designed neural network system, which achieves an increase in performance due to self-organising oscillating neural clusters. The developed neural network model is used to reduce radio deviation and associated navigation errors due to conductive obstacles or atmospheric formations.

METHODS OF REPRESENTATION OF 3D OBJECTS FOR LEARNING GENERATIVE NEURAL NETWORKS

The paper considered various methods of three-dimensional objects generating by using of neural networks. Several key elements of the methodologies of this type of synthesis of new information were singled out, on the basis of which a new way of three-dimensional objects representing was proposed for its use for typical generative models of neural networks training. The purpose of the work is to develop a method of representing of three-dimensional objects that would satisfy the criterion of high density of information useful for a generative model. The minimization of the redundancy of the information generated together with the minimization of the losses associated with the process of transition from a three-dimensional way of representing of an object to a two-dimensional one (with which the existing generative models can cope quite well) are the key aspects of the proposed method of representation. The methodology for solving of the problem given consists in building of a mathematical model of a new type of representation of three-dimensional objects; development of a software algorithm that implements a mathematical model; and testing this representation based on a typical generative neural network model. The scientific novelty is that for the first time such a type of representation of a three-dimensional object was proposed, which could be used for typical generative models training. Conclusions. The proposed method of representing of a three-dimensional object showed its viability even in the context of training of a small typical generative model DCGAN. Prospects for further research of the proposed method for training of other typical generative models were also determined, because this method could be quite easily adapted to representations of input and output data of a wide range of neural network architectures.

Modeling of limit order book data with ordered fuzzy numbers

This paper presents a novel approach to representing the Limit Order Book data at a given timestamp using the Ordered Fuzzy Numbers concept. The limit order book contains all buy and sell orders placed by investors, updated in real-time, for the most liquid securities, even several hundred times a minute. Due to its irregular nature (different and dynamic changes in the number of buy and sell orders), direct calculations on the order book data are not feasible without transforming it into feature vectors. Currently, most studies use a price level-based data representation scheme when applying deep learning models on limit order book data. However, this scheme has limitations, particularly its sensitivity to subtle perturbations that can negatively impact model performance. On the other hand, the ordered fuzzy number is a mathematical object (a pair of two functions) used to process imprecise and uncertain data. Ordered Fuzzy Numbers possess well-defined arithmetic properties. Converting the limit order book data to ordered fuzzy numbers allows the creation of a time series of ordered fuzzy numbers (order books) and use them for further calculations, e.g., to represent input data for deep learning models or employing the concept of fuzzy time series in various domains, such as defining liquidity measures based on limit order book data. In this paper, the proposed approach is tested using one-year market data from the Polish Stock Exchange for the five biggest companies. The DeepLOB model is employed to predict mid-price movement using different input data representations. The proposed representation of Limit Order Book data demonstrated remarkably stable out-of-sample prediction accuracy, even when subjected to data perturbation.

Enhancing Urban Data Analysis: Leveraging Graph-Based Convolutional Neural Networks for a Visual Semantic Decision Support System.

The persistent increase in the magnitude of urban data, combined with the broad range of sensors from which it derives in modern urban environments, poses issues including data integration, visualization, and optimal utilization. The successful selection of suitable locations for predetermined commercial activities and public utility services or the reuse of existing infrastructure arise as urban planning challenges to be addressed with the aid of the aforementioned data. In our previous work, we have integrated a multitude of publicly available real-world urban data in a visual semantic decision support environment, encompassing map-based data visualization with a visual query interface, while employing and comparing several classifiers for the selection of appropriate locations for establishing parking facilities. In the current work, we challenge the best representative of the previous approach, i.e., random forests, with convolutional neural networks (CNNs) in combination with a graph-based representation of the urban input data, relying on the same dataset to ensure comparability of the results. This approach has been inspired by the inherent visual nature of urban data and the increased capability of CNNs to classify image-based data. The experimental results reveal an improvement in several performance indices, implying a promising potential for this specific combination in decision support for urban planning problems.

O2 supplementation disambiguation in clinical narratives to support retrospective COVID-19 studies

BackgroundOxygen saturation, a key indicator of COVID-19 severity, poses challenges, especially in cases of silent hypoxemia. Electronic health records (EHRs) often contain supplemental oxygen information within clinical narratives. Streamlining patient identification based on oxygen levels is crucial for COVID-19 research, underscoring the need for automated classifiers in discharge summaries to ease the manual review burden on physicians.MethodWe analysed text lines extracted from anonymised COVID-19 patient discharge summaries in German to perform a binary classification task, differentiating patients who received oxygen supplementation and those who did not. Various machine learning (ML) algorithms, including classical ML to deep learning (DL) models, were compared. Classifier decisions were explained using Local Interpretable Model-agnostic Explanations (LIME), which visualize the model decisions.ResultClassical ML to DL models achieved comparable performance in classification, with an F-measure varying between 0.942 and 0.955, whereas the classical ML approaches were faster. Visualisation of embedding representation of input data reveals notable variations in the encoding patterns between classic and DL encoders. Furthermore, LIME explanations provide insights into the most relevant features at token level that contribute to these observed differences.ConclusionDespite a general tendency towards deep learning, these use cases show that classical approaches yield comparable results at lower computational cost. Model prediction explanations using LIME in textual and visual layouts provided a qualitative explanation for the model performance.

Optical hyperdimensional soft sensing: speckle-based touch interface and tactile sensor.

Hyperdimensional computing (HDC) is an emerging computing paradigm that exploits the distributed representation of input data in a hyperdimensional space, the dimensions of which are typically between 1,000-10,000. The hyperdimensional distributed representation enables energy-efficient, low-latency, and noise-robust computations with low-precision and basic arithmetic operations. In this study, we propose optical hyperdimensional distributed representations based on laser speckles for adaptive, efficient, and low-latency optical sensor processing. In the proposed approach, sensory information is optically mapped into a hyperdimensional space with >250,000 dimensions, enabling HDC-based cognitive processing. We use this approach for the processing of a soft-touch interface and a tactile sensor and demonstrate to achieve high accuracy of touch or tactile recognition while significantly reducing training data amount and computational burdens, compared with previous machine-learning-based sensing approaches. Furthermore, we show that this approach enables adaptive recalibration to keep high accuracy even under different conditions.

Transformers for Generalized Fast Shower Simulation

Recently, transformer-based foundation models have proven to be a generalized architecture applicable to various data modalities, ranging from text to audio and even a combination of multiple modalities. Transformers by design should accurately model the non-trivial structure of particle showers thanks to the absence of strong inductive bias, better modeling of long-range dependencies, and interpolation and extrapolation capabilities. In this paper, we explore a transformer-based generative model for detector-agnostic fast shower simulation, where the goal is to generate synthetic particle showers, i.e., the energy depositions in the calorimeter. When trained with an adequate amount and variety of showers, these models should learn better representations compared to other deep learning models, and hence should quickly adapt to new detectors. In this work, we will show the prototype of a transformer-based generative model for fast shower simulation, as well as explore certain aspects of transformer architecture such as input data representation, sequence formation, and the learning mechanism for our unconventional shower data.

Hierarchical Locality-Aware Deep Dictionary Learning for Classification

Deep dictionary learning (DDL) shows good performance in visual classification tasks. However, almost all existing DDL methods ignore the locality relationships between the input data representations and the learned dictionary atoms, and learn sub-optimal representations in the feature coding stage, which are less conducive to classification. To this end, we propose a hierarchical locality-aware deep dictionary learning (HILADLE) framework for classification, which can learn locality-constrained dictionaries at different abstract levels through hierarchical dictionary learning. The locality constraints play an important role in learning informative dictionary atoms while preserving the data structure in the original input feature space. Moreover, instead of using an identity activation function like existing DDL methods, we further boost the generalization performance of our HILADLE method with a ReLU activation function to deal with the overfitting issue caused by over-parameterization, inspired by its effectiveness in deep neural networks. Finally, the concatenation of all feature representations learned at different layers is used as input to the final classifier. We demonstrate, through an extensive set of experiments on several benchmark face recognition, image classification, and age estimation datasets, that our method is able to surpass several dictionary learning, deep dictionary learning and deep learning methods.

Detect malicious websites by building a neural network to capture global and local features of websites

With the development of the digital age, the Internet has become an integral part of our daily lives. However, it has also brought about a series of security challenges, among which malicious websites are particularly prominent. These websites often lure ignorant users by disguising themselves as legitimate services or through various fraudulent means to commit identity theft, distribute malware, or launch other forms of cyberattacks. Therefore the detection of malicious websites is very necessary. Traditionally, many malicious website detection methods rely on machine learning techniques, some of which require manual extraction of features, which may result in a time-consuming prediction process. Despite the existence of machine learning models that can automatically extract features, including unsupervised ones, capturing the subtleties of malicious website features is still a challenge. In recent years, deep learning has been gaining attention as a method for automated feature learning. It is capable of capturing and understanding the content of a website in greater depth, thus making classification and detection more accurate and efficient. Although deep learning shows its potential in capturing advanced features, its performance depends on the input data and the chosen model architecture. Both efficiently constructing feature representations of input data and building efficient model architectures to capture features are currently major challenges. For this reason, we propose a new approach for malicious website detection. This method uses wordpiece-level features to represent the information of malicious websites. Combination of multi-filter text convolutional neural network and multi-head self-attention mechanism is used for model construction. This enables the model to capture both global and local features of the input data. Compared to common deep learning methods, our approach captures the features of malicious websites better.

Hybrid variational autoencoder for time series forecasting

Variational autoencoders (VAE) are powerful generative models that learn the latent representations of input data as random variables. Recent studies show that VAE can flexibly learn the complex temporal dynamics of time series and achieve more promising forecasting results than deterministic models. However, a major limitation of existing works is that they fail to jointly learn the local patterns (e.g., seasonality and trend) and temporal dynamics of time series for forecasting. Accordingly, we propose a novel hybrid variational autoencoder (HyVAE) to integrate the learning of local patterns and temporal dynamics by variational inference for time series forecasting. Experimental results on four real-world datasets show that the proposed HyVAE achieves better forecasting results than various counterpart methods, as well as two HyVAE variants that only learn the local patterns or temporal dynamics of time series, respectively.

Detection of valvular heart diseases combining orthogonal non-negative matrix factorization and convolutional neural networks in PCG signals

Background and objective:Valvular heart disease (VHD) is associated with elevated mortality rates. Although transthoracic echocardiography (TTE) is the gold standard detection tool, phonocardiography (PCG) could be an alternative as it is a cost-effective and noninvasive method for cardiac auscultation. Many researchers have dedicated their efforts to improving the decision-making process and developing robust and precise approaches to assist physicians in providing reliable diagnoses of VHD. Methods:This research proposes a novel approach for the detection of anomalous valvular heart sounds from PCG signals. The proposed approach combines orthogonal non-negative matrix factorization (ONMF) and convolutional neural network (CNN) architectures in a three-stage cascade. The aim of the proposal is to improve the learning process by identifying the optimal ONMF temporal or spectral patterns for accurate detection. In the first stage, the time–frequency representation of the input PCG signal is computed. Next, band-pass filtering is performed to locate the spectral range that is most relevant for the presence of such cardiac abnormalities. In the second stage, the temporal and spectral cardiac structures are extracted using the ONMF approach. These structures are utilized in the third stage and fed into the CNN architecture to detect abnormal heart sounds. Results:Several state-of-the-art CNN architectures, such as LeNet5, AlexNet, ResNet50, VGG16 and GoogLeNet, have been evaluated to determine the effectiveness of using ONMF temporal features for VHD detection. The results reveal that the integration of ONMF temporal features with a CNN classifier significantly improve VHD detection. Specifically, the proposed approach achieves an accuracy improvement of approximately 45% when ONMF spectral features are used and 35% when time–frequency features from the short-time Fourier transform (STFT) spectrogram are used. Additionally, feeding ONMF temporal features into low-complexity CNN architectures yields competitive results comparable to those obtained with complex architectures. Conclusions:The temporal structure factorized by ONMF plays a critical role in distinguishing between normal heart sounds and abnormal heart sounds since the repeatability of normal heart cycles is disrupted by the presence of cardiac abnormalities. Consequently, the results highlight the importance of appropriate input data representation in the learning process of CNN models in the biomedical field of valvular heart sound detection.

Experimental Observations of the Topology of Convolutional Neural Network Activations

Topological data analysis (TDA) is a branch of computational mathematics, bridging algebraic topology and data science, that provides compact, noise-robust representations of complex structures. Deep neural networks (DNNs) learn millions of parameters associated with a series of transformations defined by the model architecture resulting in high-dimensional, difficult to interpret internal representations of input data. As DNNs become more ubiquitous across multiple sectors of our society, there is increasing recognition that mathematical methods are needed to aid analysts, researchers, and practitioners in understanding and interpreting how these models' internal representations relate to the final classification. In this paper we apply cutting edge techniques from TDA with the goal of gaining insight towards interpretability of convolutional neural networks used for image classification. We use two common TDA approaches to explore several methods for modeling hidden layer activations as high-dimensional point clouds, and provide experimental evidence that these point clouds capture valuable structural information about the model's process. First, we demonstrate that a distance metric based on persistent homology can be used to quantify meaningful differences between layers and discuss these distances in the broader context of existing representational similarity metrics for neural network interpretability. Second, we show that a mapper graph can provide semantic insight as to how these models organize hierarchical class knowledge at each layer. These observations demonstrate that TDA is a useful tool to help deep learning practitioners unlock the hidden structures of their models.

Biologically informed variational autoencoders allow predictive modeling of genetic and drug-induced perturbations.

Variational autoencoders (VAEs) have rapidly increased in popularity in biological applications and have already successfully been used on many omic datasets. Their latent space provides a low-dimensional representation of input data, and VAEs have been applied, e.g. for clustering of single-cell transcriptomic data. However, due to their non-linear nature, the patterns that VAEs learn in the latent space remain obscure. Hence, the lower-dimensional data embedding cannot directly be related to input features. To shed light on the inner workings of VAE and enable direct interpretability of the model through its structure, we designed a novel VAE, OntoVAE (Ontology guided VAE) that can incorporate any ontology in its latent space and decoder part and, thus, provide pathway or phenotype activities for the ontology terms. In this work, we demonstrate that OntoVAE can be applied in the context of predictive modeling and show its ability to predict the effects of genetic or drug-induced perturbations using different ontologies and both, bulk and single-cell transcriptomic datasets. Finally, we provide a flexible framework, which can be easily adapted to any ontology and dataset. OntoVAE is available as a python package under https://github.com/hdsu-bioquant/onto-vae.

Cluster analysis of autoencoder-extracted FDG PET/CT features identifies multiple myeloma patients with poor prognosis

F-18 fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT) is a robust imaging modality used for staging multiple myeloma (MM) and assessing treatment responses. Herein, we extracted features from the FDG PET/CT images of MM patients using an artificial intelligence autoencoder algorithm that constructs a compressed representation of input data. We then evaluated the prognostic value of the image-feature clusters thus extracted. Conventional image parameters including metabolic tumor volume (MTV) were measured on volumes-of-interests (VOIs) covering only the bones. Features were extracted with the autoencoder algorithm on bone-covering VOIs. Supervised and unsupervised clustering were performed on image features. Survival analyses for progression-free survival (PFS) were performed for conventional parameters and clusters. In result, supervised and unsupervised clustering of the image features grouped the subjects into three clusters (A, B, and C). In multivariable Cox regression analysis, unsupervised cluster C, supervised cluster C, and high MTV were significant independent predictors of worse PFS. Supervised and unsupervised cluster analyses of image features extracted from FDG PET/CT scans of MM patients by an autoencoder allowed significant and independent prediction of worse PFS. Therefore, artificial intelligence algorithm–based cluster analyses of FDG PET/CT images could be useful for MM risk stratification.

Exploiting device-level non-idealities for adversarial attacks on ReRAM-based neural networks

Resistive memory (ReRAM) or memristor devices offer the prospect of more efficient computing. While memristors have been used for a variety of computing systems, their usage has gained significant popularity in the domain of deep learning. Weight matrices in deep neural networks can be mapped to crossbar architectures with memristive junctions, generally resulting in superior performance and energy efficiency. However, the nascent nature of ReRAM technology is directly associated with the presence of inherent non-idealities in the ReRAM devices currently available. Deep neural networks have already been shown to be susceptible to adversarial attacks, often by targeting vulnerabilities in the networks’ internal representation of input data. In this paper, we explore the causal relationship between device-level non-idealities in ReRAM devices and the classification performance of memristor-based neural network accelerators. Specifically, our aim is to generate images which bypass adversarial defense mechanisms in software neural networks but trigger non-trivial performance discrepancies in ReRAM-based neural networks. To this end, we have proposed a framework to generate adversarial images in the hypervolume between the two decision boundaries, thereby leveraging non-ideal device behavior for performance detriment. We employ state-of-the-art tools in explainable artificial intelligence to characterize our adversarial image samples, and derive a new metric to quantify susceptibility to adversarial attacks at the pixel and device-levels.

Deep feature screening: Feature selection for ultra high-dimensional data via deep neural networks

The applications of traditional statistical feature selection methods to high-dimension, low-sample-size data often struggle and encounter challenging problems, such as overfitting, curse of dimensionality, computational infeasibility, and strong model assumptions. In this paper, we propose a novel two-step nonparametric approach called Deep Feature Screening (DeepFS) that can overcome these problems and identify significant features with high precision for ultra high-dimensional, low-sample-size data. This approach first extracts a low-dimensional representation of input data and then applies feature screening on the original input feature space based on multivariate rank distance correlation recently developed by Deb and Sen (2021). This approach combines the strengths of both deep neural networks and feature screening, thereby having the following appealing features in addition to its ability of handling ultra high-dimensional data with small number of samples: (1) it is model free and distribution free; (2) it can be used for both supervised and unsupervised feature selection; and (3) it is capable of recovering the original input data. The superiority of DeepFS is demonstrated via extensive simulation studies and real data analyses.