Layer Perceptron Research Articles

Cyber Hacking Breaches Detection Using Machine Learning

Cyber hacking breach prediction is one of the emerging technologies and detecting and predicting breaches through computer algorithms has become a very challenging task To make malware detection more effective, scalable and more efficient than traditional system calls human involvement Machine learning to be used for breach detection and prediction The main goal is a series of cyber hacking attacks each of which will harm the person’s information and financial reputation. Government and non-profit organizations’ data, such as user and company information,, can be compromised, posing a risk to their finances and reputation if they collect information from websites and social networks it can trigger a cyber attack. Organizations such as the healthcare sector are capable of holding sensitive information that must be handled discreetly and securely. Data breaches can lead to identity theft, fraud, and other losses. The findings show that 70% of breaches affect a wide range of organisations, including healthcare providers. The investigation indicates a possible data breach. Due to the heavy usage of computer programs and security on the host and network, there is a risk of data breach. Machine learning can be used to detect these attacks. Research uses machine learning models to protect against web security flaws. The data set is available from the Privacy Rights Clearing House. Teaching employees how to use modern security measures can reduce data breaches. This can help to understand attack detection and data security. Machine learning models such as Random Forest, Decision Tree, k-means and Multi- layer Perceptron are used to predict data violations

StochCA: A novel approach for exploiting pretrained models with cross-attention

Utilizing large-scale pretrained models is a well-known strategy to enhance performance on various target tasks. It is typically achieved through fine-tuning pretrained models on target tasks. However, naï ve fine-tuning may not fully leverage knowledge embedded in pretrained models. In this study, we introduce a novel fine-tuning method, called stochastic cross-attention (StochCA), specific to Transformer architectures. This method modifies the Transformer’s self-attention mechanism to selectively utilize knowledge from pretrained models during fine-tuning. Specifically, in each block, instead of self-attention, cross-attention is performed stochastically according to the predefined probability, where keys and values are extracted from the corresponding block of a pretrained model. By doing so, queries and channel-mixing multi-layer perceptron layers of a target model are fine-tuned to target tasks to learn how to effectively exploit rich representations of pretrained models. To verify the effectiveness of StochCA, extensive experiments are conducted on benchmarks in the areas of transfer learning and domain generalization, where the exploitation of pretrained models is critical. Our experimental results show the superiority of StochCA over state-of-the-art approaches in both areas. Furthermore, we demonstrate that StochCA is complementary to existing approaches, i.e., it can be combined with them to further improve performance. We release the code at https://github.com/daintlab/stochastic_cross_attention.

Fault diagnosis of induction motor in the cooling water supply system using a multi-channel data fusion transformer with limited sample conditions*

Abstract Induction motors play a vital role in the cooling water supply system of hydropower facilities. However, it is not feasible to collect sufficient fault samples in a hydropower station. The scarcity of labeled samples poses a challenge in developing powerful diagnostic models with high classification accuracy. To address this challenge, this paper proposes a multi-channel data fusion strategy based on a transformer for feature enhancement. Initially, the original signals are transferred into non-overlapping single-channel data patches to preserve correlation features across different channels. Next, temporal and spatial attention modules are applied to process the data patches, which can learn and fuse temporal and spatial information, respectively. Subsequently, the data patches are embedded to retain position information and represent fault-related features through class embedding, which are further processed by a transformer encoder with self-attention mechanisms. Finally, the classification task is achieved by using a multilayer perceptron layer connected to the class embedding. While dealing with limited training samples, the proposed method can learn robust features that are beneficial to improve the fault diagnosis ability of induction motors. The comparison of the proposed method with three basic models and two improved methods demonstrates the superiority of the proposed method in accuracy and feature clustering performance under limited sample conditions. In addition, ablation experiments demonstrate the necessity of each module in the proposed method.

Improving multi-site photovoltaic forecasting with relevance amplification: DeepFEDformer-based approach

The present research on photovoltaic (PV) forecasting is devoted to the use of spatial information about PV sites to improve the accuracy of the models, but most of the models have to increase their network complexity to learn the spatial dependence. In this paper, we propose to take advantage of the known geographic location information of PV sites and embed them directly into the input information of Decoder, which makes it easier for the model to focus its attention. We refer to this process as relevance amplification. Based on this, this paper proposes Relevance Amplification based DeepFEDformer (RAD-FEDformer), where DeepFEDformer adds multiple Multi-Layer Perceptron (MLP) layers to FEDformer to improve the perception of deep features. The Relevance Amplification Module (RAM) is designed to receive geographic correlation information as a way to enhance its influence on the Seasonality component of the Decoder input and improve the performance of the attention mechanism in the model. Using the power generation data from PV plants distributed in 11 regions of Belgium as our case study, we evaluated the performance of RAD-FEDformer in predicting PV data at 48/96/192 time steps into the future for each plant. Compared to other Transformer family models, RAD-FEDformer achieved SOTA results by demonstrating significant R2 improvements of 16.70%, 51.45%, and 23.50% over FEDformer, along with an average MSE reduction of 16.9%. We designed Ablation Experiments to validate and compare the performance of MLPs of different sizes based on the ETTm1 dataset and discussed the performance of RAM and its effect on the attention mechanism based on the PV dataset. The results show that our model enhances the performance of the attention mechanism in Multi-Site PV prediction scenarios, and conclude that the optimization of RAM is more effective in longer sequence predictions.

CSDFormer: A cloud and shadow detection method for landsat images based on transformer

Cloud and shadow (CS) detection is crucial prerequisite for application of remote sensing images. Current deep learning-based detection algorithms mainly employ Convolutional Neural Networks (CNNs). However, the local receptive field in CNNs cannot effectively capture global contextual information, which hinders accurate characterization of the dependency between clouds and shadows. In vision Transformers, self-attention mechanisms can effectively capture the long-distance dependencies between different regions in an image. Inspired by this, this paper proposed a new CS Detection algorithm based on a Transformer, called CSDFormer. Specifically, we exclusively employed a hierarchical Transformer structure in the encoder stage to extract features of CS. Each Transformer layer contains several multi-head self-attention mechanisms for calculating pixel-wise long-distance connectivity. The designed structure enables the Transformer to better extract global context information, which helps to strengthen the comprehension of the semantic relationships between clouds and shadows. Benefiting from the global feature extraction capability of the encoder stage, we employed several simple multilayer perceptron layers for multi-scale feature map fusion and pixel classification in the decoder stage. The proposed CSDFormer was validated using 898 Landsat 8 Biome images with 512 × 512 pixels, producing an overall accuracy of 95.28 % and a mean intersection over union of 84.08 %, outperforming three state-of-the-art CNN-based algorithms. CSDFormer is consistently more accurate in detection of both clouds and shadows. Owing to the parallel computing capability of the self-attention mechanism, CSDFormer is computationally more efficient than the three CNN-based benchmark methods. For the input spectral bands, the performance of CSDFormer produced can be further enhanced with additional thermal infrared bands.

Spiking NeRF: Representing the Real-World Geometry by a Discontinuous Representation

A crucial reason for the success of existing NeRF-based methods is to build a neural density field for the geometry representation via multiple perceptron layers (MLPs). MLPs are continuous functions, however, real geometry or density field is frequently discontinuous at the interface between the air and the surface. Such a contrary brings the problem of unfaithful geometry representation. To this end, this paper proposes spiking NeRF, which leverages spiking neurons and a hybrid Artificial Neural Network (ANN)-Spiking Neural Network (SNN) framework to build a discontinuous density field for faithful geometry representation. Specifically, we first demonstrate the reason why continuous density fields will bring inaccuracy. Then, we propose to use the spiking neurons to build a discontinuous density field. We conduct a comprehensive analysis for the problem of existing spiking neuron models and then provide the numerical relationship between the parameter of the spiking neuron and the theoretical accuracy of geometry. Based on this, we propose a bounded spiking neuron to build the discontinuous density field. Our method achieves SOTA performance. The source code and the supplementary material are available at https://github.com/liaozhanfeng/Spiking-NeRF.

A deep learning algorithm model to automatically score and grade obstructive sleep apnea in adult polysomnography.

Polysomnography (PSG) is unique in diagnosing sleep disorders, notably obstructive sleep apnea (OSA). Despite its advantages, manual PSG data grading is time-consuming and laborious. Thus, this research evaluated a deep learning-based automated scoring system for respiratory events in sleep-disordered breathing patients. A total of 1000 case PSG data were enrolled to develop a deep learning algorithm. Of the 1000 data, 700 were distributed for training, 200 for validation, and 100 for testing. The respiratory events scoring deep learning model is composed of five sequential layers: an initial layer of perceptrons, followed by three consecutive layers of long short-term memory cells, and ultimately, an additional two layers of perceptrons. The PSG data of 100 patients (simple snoring, mild, moderate, and severe OSA; n = 25 in each group) were selected for validation and testing of the deep learning model. The algorithm demonstrated high sensitivity (95% CI: 98.06-98.51) and specificity (95% CI: 95.46-97.79) across all OSA severities in detecting apnea/hypopnea events, compared to manual PSG analysis. The deep learning model's area under the curve values for predicting OSA in apnea-hypopnea index ≥ 5, 15, and 30 groups were 0.9402, 0.9388, and 0.9442, respectively, showing no significant differences between each group. The deep learning algorithm employed in our study showed high accuracy in identifying apnea/hypopnea episodes and assessing the severity of OSA, suggesting the potential for enhancing both the efficiency and accuracy of automated respiratory event scoring in PSG through advanced deep learning techniques.

Characterizing Internet Card User Portraits for Efficient Churn Prediction Model Design

Cellular Internet card (IC) as a new business model emerges, which penetrates rapidly and holds the potential to foster a great business market. However, with the explosive growth of IC users, the user churn problem becomes severe, affecting the IC business significantly, while there is lacking appropriate techniques in the literature to deal with the issue. In this paper, we take the lead to study one large-scale data set from a provincial network operator of China, which contains about 4 million IC users and 22 million traditional card (TC) users. We first justify the IC user churn issue with data, and categorize the user churning reasons. Then, we shed light on understanding user portraits, which is the building block to enable efficient model design. Particularly, we conduct a systematical analytics on usage data by studying the difference of two types of users, examining the impact of user properties, and characterizing the user Internet using behaviors. Finally, by using the IC user portraits and usage patterns, we propose an <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">IC</u> user <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C</u> hurn <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">P</u> rediction model, named <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ICCP</i> , which consists of a feature extraction component and a learning-based churn prediction architecture design. For feature extraction, both the static portrait features and temporal sequential features are captured. In the learning architecture, we devise the principal component analysis (PCA) block and the embedding/transformer layers to learn the respective information of two types of features, which are collectively fed into the classification multilayer perceptron layer (MPL) for churn prediction. A reference implementation of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ICCP</i> is conducted within the telecom system and extensive experiments corroborate the efficiency of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ICCP</i> .

Hybrid carbon price forecasting using a deep augmented FEDformer model and multimodel optimization piecewise error correction

Accurate and robust carbon price forecasting is crucial for stabilizing the carbon financial market and reducing reliance on fossil resources. This study introduces a novel hybrid model, combining a deep augmented frequency enhanced decomposed transformer (DA-FEDformer), improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), and a multimodel optimization piecewise error correction, for accurate carbon price prediction. To leverage frequency domain information, the proposed DA-FEDformer model enhances the decoder and encoder layers by incorporating a multilayer perceptron layer. A novel improved kernel mean square error loss function is devised for the model, and the evolved sign momentum optimizer is employed for further optimization. To further improve the prediction accuracy, a data postprocessing method is integrated into the model. The DA-FEDformer predicts the carbon price, and the ICEEMDAN method decomposes the error sequence into subsequences. These subsequences are forecasted using five different models. Finally, the DA-FEDformer prediction results and the error correction results are combined to obtain the final carbon price sequence. This hybrid model is systematically evaluated using carbon price data from Hubei, Guangzhou, and Beijing in China. The results demonstrate that the hybrid model achieves favorable performance on three datasets. This innovative approach provides a promising solution for carbon price forecasting in China.

Disentangle interest trend and diversity for sequential recommendation

Based on historical behaviors, sequential recommendation endeavors to predict what a user prefers next. The recent efforts are mainly devoted to modeling the user’s interests evolution process or mining multi-interests for recommendation. However, it is largely overlooked that the interest trend (i.e., the evolution of the main interest) and the interest diversity (i.e., the scattered potential interests) could complement each other for better performance. Specifically, the interest trend reveals the user’s basic interest and its evolution, which is satisfied by similarity recommendations. Nevertheless, interest diversity covers the various interests caused by some external environmental influence, e.g., fashion trends and advertisements, exploring users’ potential interests or interest diversity will facilitate the model for diversity and serendipity recommendation. In a way, these two factors have conflicting aims, we argue that they should be disentangled in modeling first and recombined when making personalized recommendation.To instantiate this idea, we propose a simple yet effective model, dubbed Teddy (disentangles interest trend and diversity), which disentangles and then jointly models the aforementioned two factors under a unified framework. Particularly, in Teddy, an adaptive masking mechanism is first introduced to split the user’s historical items into two parts revealing her major interest trend and scattered interest diversity respectively for interests disentanglement. Then, a temporal convolutional network (TCN) is utilized to capture the evolution process of the user’s major interest trend. For the scattered interest diversity modeling, we further choose to apply Multilayer Perceptron (MLP) layers with max-pooling mechanism to extract the significant or dominant preference signals. The predicted scores generated by these two modules are aggregated together to integrate both interest trend and interest diversity for the final recommendation. Extensive experiments over four public datasets demonstrate the superiority of our proposed Teddy against a series of SOTA alternatives on the benchmark metrics.

Learning Signed Hyper Surfaces for Oriented Point Cloud Normal Estimation.

We propose a novel method called SHS-Net for point cloud normal estimation by learning signed hyper surfaces, which can accurately predict normals with global consistent orientation from various point clouds. Almost all existing methods estimate oriented normals through a two-stage pipeline, i.e., unoriented normal estimation and normal orientation, and each step is implemented by a separate algorithm. However, previous methods are sensitive to parameter settings, resulting in poor results from point clouds with noise, density variations and complex geometries. In this work, we introduce signed hyper surfaces (SHS), which are parameterized by multi-layer perceptron (MLP) layers, to learn to estimate oriented normals from point clouds in an end-to-end manner. The signed hyper surfaces are implicitly learned in a high-dimensional feature space where the local and global information is aggregated. Specifically, we introduce a patch encoding module and a shape encoding module to encode a 3D point cloud into a local latent code and a global latent code, respectively. Then, an attention-weighted normal prediction module is proposed as a decoder, which takes the local and global latent codes as input to predict oriented normals. Experimental results show that our algorithm outperforms the state-of-the-art methods in both unoriented and oriented normal estimation.

Multichannel Orthogonal Transform-Based Perceptron Layers for Efficient ResNets.

In this article, we propose a set of transform-based neural network layers as an alternative to the 3 x 3 Conv2D layers in convolutional neural networks (CNNs). The proposed layers can be implemented based on orthogonal transforms, such as the discrete cosine transform (DCT), Hadamard transform (HT), and biorthogonal block wavelet transform (BWT). Furthermore, by taking advantage of the convolution theorems, convolutional filtering operations are performed in the transform domain using elementwise multiplications. Trainable soft-thresholding layers, that remove noise in the transform domain, bring nonlinearity to the transform domain layers. Compared with the Conv2D layer, which is spatial-agnostic and channel-specific, the proposed layers are location-specific and channel-specific. Moreover, these proposed layers reduce the number of parameters and multiplications significantly while improving the accuracy results of regular ResNets on the ImageNet-1K classification task. Furthermore, they can be inserted with a batch normalization (BN) layer before the global average pooling layer in the conventional ResNets as an additional layer to improve classification accuracy.

Design of spiking neural network architecture based on dendritic computation principles

The paper presents the hardware architecture design of a spiking neural network (SNN) based on dendritic computation principles. The integration of active dendritic properties into the neuronal structure of SNN aims to minimize the number of functional blocks required for hardware implementation, including synaptic connections and neurons. The available memory on the neuromorphic architecture imposes limitations on implementation, hence the need to reduce the number of functional blocks. As a test task for the SNN based on dendritic computations, we selected the image classification of eight symbols, consisting of digits one through eight. These symbols are depicted as 3×7 pixel, 1-bit images. Active dendritic properties were analyzed using the “delay plasticity” [1] principle, which introduces the mechanism of adjusting input signal delays in spiking neuron inputs. We designed an SNN model with complementary delay inputs, referred to as the active dendrite SNN, as a principle implementation. Input spikes arriving at the primary inputs are duplicated to the delay inputs after a modifiable time delay. For convenience, each delay input was set at a single value. The input images were scanned sequentially. The neural network received three direct and three inverse inputs from the six main inputs that were coded with spikes corresponding to three pixels of a string. An “on” pixel was coded with a spike arriving at a direct input, while an “off” pixel was coded with a spike arriving at the corresponding inverse input. The line scanning time was 10 μs, input width was 1 μs, and delay time was 5 μs. The optimization of spiking neuron parameters was performed through a stochastic search algorithm based on simulated annealing. The parameters optimized for the Leaky-Integrate-and-Fire (LIF) neurons included the leakage time constant (22.8 μs), firing threshold (1150 arbitrary units), and refractory period (1 μs). The active dendrite SNN training employed the tempotron learning rule [2]. The training optimized the following parameters: the maximum change in synaptic weight on potentiation and depression (0.7 and –3 arbitrary units, respectively) and the synaptic weight’s upper bound (195 arbitrary units). Complementary delayed inputs facilitated the learning of the order in which input patterns arrived for SNN neurons during training. The paper compares an SNN architecture based on dendritic computations to our previously designed two-layer SNN with a hidden perceptron layer and an output layer consisting of LIF neurons [3]. Using the same LIF neuron design, input image coding, and LIF neuron layer structure as in the proposed architecture, our two-layer SNN with a hidden perceptron layer and output layer of LIF neurons successfully recognized 3×5 images of three symbols with only 10 neurons and 63 synapses. Alternatively, the active dendrite SNN was able to recognize 3×7 images of eight symbols with four neurons and 48 synaptic weights. In conclusion, incorporating active dendrite properties into the SNN architecture for image recognition resulted in optimized functional block usage, lowering the number of neurons and synapses used by 60 and 24%, respectively.

Enhancing Fashion Classification with Vision Transformer (ViT) and Developing Recommendation Fashion Systems Using DINOVA2

As e-commerce platforms grow, consumers increasingly purchase clothes online; however, they often need clarification on clothing choices. Consumers and stores interact through the clothing recommendation system. A recommendation system can help customers to find clothing that they are interested in and can improve turnover. This work has two main goals: enhancing fashion classification and developing a fashion recommendation system. The main objective of fashion classification is to apply a Vision Transformer (ViT) to enhance performance. ViT is a set of transformer blocks; each transformer block consists of two layers: a multi-head self-attention layer and a multilayer perceptron (MLP) layer. The hyperparameters of ViT are configured based on the fashion images dataset. CNN models have different layers, including multi-convolutional layers, multi-max pooling layers, multi-dropout layers, multi-fully connected layers, and batch normalization layers. Furthermore, ViT is compared with different models, i.e., deep CNN models, VGG16, DenseNet-121, Mobilenet, and ResNet50, using different evaluation methods and two fashion image datasets. The ViT model performs the best on the Fashion-MNIST dataset (accuracy = 95.25, precision = 95.20, recall = 95.25, F1-score = 95.20). ViT records the highest performance compared to other models in the fashion product dataset (accuracy = 98.53, precision = 98.42, recall = 98.53, F1-score = 98.46). A recommendation fashion system is developed using Learning Robust Visual Features without Supervision (DINOv2) and a nearest neighbor search that is built in the FAISS library to obtain the top five similarity results for specific images.

Multi-layer parallel transformer model for detecting product quality issues and locating anomalies based on multiple time‑series process data in Industry 4.0

Smart manufacturing systems typically consist of multiple machines with different processing durations. The continuous monitoring of these machines produces multiple time-series process data (MTPD), which have four characteristics: low data value density, diverse data dimensions, transmissible processing states, and complex coupling relationships. Using MTPD for product quality issue detection and rapid anomaly location can help dynamically adjust the control of smart manufacturing systems and improve manufacturing yield. This study proposes a multi-layer parallel transformer (MLPT) model for product quality issue detection and rapid anomaly location in Industry 4.0, based on proper modeling of the MTPD of smart manufacturing systems. The MLPT consists of multiple customized encoder models that correspond to the machines, each using a customized partition strategy to determine the token size. All encoders are integrated in parallel and output to the global multi-layer perceptron layer, which improves the accuracy of product quality issue detection and simultaneously locates anomalies (including key time steps and key sensor parameters) in smart manufacturing systems. An empirical study was conducted on a fan-out, panel-level package (FOPLP) production line. The experimental results show that the MLPT model can detect product quality issues more accurately than other methods. It can also rapidly realize anomalous locations in smart manufacturing systems.

Multi-objective optimization of thermophysical properties GO powders-DW/EG Nf by RSM, NSGA-II, ANN, MLP and ML

In this study, prediction, modeling, and optimization have been performed for four TPH properties of graphene oxide nano powder-deionized water/ethylene glycol nf, which is unique compared to other studies. Response surface methodology, artificial neural networks based on multiple layers of perceptron, and algorithms based on machine learning have been developed for prediction and modeling. RSM modeling resulted in coefficients of determination of 0.9984, 0.9986, 0.9995, and 0.9996 for TC (k), density (ρ), SHC (cp), and viscosity (μ), respectively. The highest prediction errors for RSM models were 0.3644%, 0.0374%, 2.049%, and 0.2296% for k, ρ, μ, and cp. A higher temperature and a higher WF of NPs increased the TC of the nf. The maximum MLP model error was 0.43%, 6.59%, 12.64%, and 1.04% for ρ, cp, μ, and k, respectively. TC and SHC were optimized using the NSGA-II algorithm. The NSGA-II procedure indicated the maximum k and cp occurred at the highest temperatures. The temperature must be kept at its maximum to reach the optimal stage. Also, it was proven that temperature is a much more significant parameter than the nanoparticle WF.

SedSVD: Statement-level software vulnerability detection based on Relational Graph Convolutional Network with subgraph embedding

Current deep-learning based vulnerability detection methods have been proven more automatic and correct to a certain extent, nonetheless, they are limited to detect at function-level or file-level, which can hinder software developers from acquiring more detailed information and conducting more targeted repairs. Graph-based detection methods have shown dominant performance over others. Unfortunately, the information they reveal has not been fully utilized. We design SedSVD (Subgraph embedding driven Statement-level Vulnerability Detection) with two objectives: (i) to better utilize the information the code-related graphs can reflect; (ii) to detect vulnerabilities at a finer-grained level. In our work, we propose a novel graph-based detection framework that embeds graphs at subgraph-level to realize statement-level detection. It first leverages Code Property Graph (CPG) to learn both semantic and syntactic information from source code, and then selects several center nodes (code elements) in CPG to build their subgraphs. After embedding each subgraph with its nodes and edges, we apply Relational Graph Convolutional Network (RGCN) to process different edges differently. A Multi-Layer Perceptron (MLP) layer is further added to ensure its prediction performance. We conduct our experiments on C/C++ projects from NVD and SARD. Experimental results show that SedSVD achieves 95.15% in F1-measure which proves our work to be more effective. Our work detects at a finer-grained level and achieves higher F1-measure than existing state-of-art vulnerability detection techniques. Besides, we provide a more detailed detection report pointing the specific error code elements within statements.

0945 A Robust Hybrid algorithm for automatic respiratory events scoring in adults

Abstract Introduction Even though Polysomnography has many advantages, manual scoring is a labor intensive and time consuming. Therefore, automatic scoring system using a deep learning algorithm has been developed and commercialized for clinical sites. However, it is challenging to explain the results of respiratory events and to measure a precise event’s duration. To overcome these challenges, a new hybrid algorithm was proposed to explain why this event was categorized and provided the accurate duration of respiratory events. Methods In order to classify respiratory events - OSA, MSA, CSA, O-hypopnea, and C-hypopnea, the proposed algorithm consists of two algorithms – deep learning and post processing algorithms. Deep learning algorithm was used to detect candidates of respiratory events and post processing algorithm was used to determine the final classification and to measure duration of events. Deep learning model consisted of five layers: a perceptron layer, three consecutive layers of long short-term memory (LSTM) cells, and two more perceptron layers. A total of 1,000 case PSG data were enrolled to develop this algorithm. To classify respiratory events from the results provided by the deep learning algorithm, auto-correlation, peak detection, de-noise filtering and transition detection were used to obtain respiratory rate and measure signal amplitude. Using these algorithm’s results, amplitude and time of baseline breathing and respiratory events were calculated and classified based on the AASM’s guidelines. Results To evaluate the performance of proposed algorithm, 30 PSG data were selected and compared with the results of deep learning algorithm. The test dataset consisted of thirty subjects with 15 OSA, 5 PLMS, 2 insomnia and 8 healthy. The kappa value of deep learning algorithms showed 0.77 (0.759, 0.78) and that of proposed algorithm was 0.838 (0.832, 0.843). They showed better agreement by 0.068 than that of deep learning algorithm. In addition, the proposed algorithm provided the classification results such as OSA, MSA, CSA, O-hypopnea and C-hypopnea and explanations of the reason of event detection. Conclusion The proposed algorithm showed better agreement and versatile classified results. This algorithm will be adopted for new automatic sleep scoring solution and this solution will be very helpful to use in real clinical sites. Support (if any)

Web-Informed-Augmented Fake News Detection Model Using Stacked Layers of Convolutional Neural Network and Deep Autoencoder

Today, fake news is a growing concern due to its devastating impacts on communities. The rise of social media, which many users consider the main source of news, has exacerbated this issue because individuals can easily disseminate fake news more quickly and inexpensive with fewer checks and filters than traditional news media. Numerous approaches have been explored to automate the detection and prevent the spread of fake news. However, achieving accurate detection requires addressing two crucial aspects: obtaining the representative features of effective news and designing an appropriate model. Most of the existing solutions rely solely on content-based features that are insufficient and overlapping. Moreover, most of the models used for classification are constructed with the concept of a dense features vector unsuitable for short news sentences. To address this problem, this study proposed a Web-Informed-Augmented Fake News Detection Model using Stacked Layers of Convolutional Neural Network and Deep Autoencoder called ICNN-AEN-DM. The augmented information is gathered from web searches from trusted sources to either support or reject the claims in the news content. Then staked layers of CNN with a deep autoencoder were constructed to train a probabilistic deep learning-base classifier. The probabilistic outputs of the stacked layers were used to train decision-making by staking multilayer perceptron (MLP) layers to the probabilistic deep learning layers. The results based on extensive experiments challenging datasets show that the proposed model performs better than the related work models. It achieves 26.6% and 8% improvement in detection accuracy and overall detection performance, respectively. Such achievements are promising for reducing the negative impacts of fake news on communities.

Node classification with graph neural network based centrality measures and feature selection

&lt;span lang="EN-US"&gt;Graph neural networks (GNNs) are a new topic of research in data science where data structure graphs are used as important components for developing and training neural networks. GNN always learns the weight importance of the neighbor for perform message aggregation in which the feature vectors of all neighbors are aggregated without considering whether the features are useful or not. Using such more informative features positively affect the performance of the GNN model. So, in this paper i) after selecting a subset of features to define important node features, we present new graph features’ explanation methods based on graph centrality measures to capture rich information and determine the most important node in a network. Through our experiments, we find that selecting certain subsets of these features and adding other features based on centrality measure can lead to better performance across a variety of datasets and ii) we introduce a major design strategy for graph neural networks. Specifically, we suggest using batch renormalization as normalization over GNN layers. Combining these techniques, representing features based on centrality measures that passed to multilayer perceptron (MLP) layer which is then passed to adjusted GNN layer, the proposed model achieves greater accuracy than modern GNN models.&lt;/span&gt;