Multilayer Classifier Research Articles

Deep Hierarchical Multimodal Metric Learning.

Multimodal metric learning aims to transform heterogeneous data into a common subspace where cross-modal similarity computing can be directly performed and has received much attention in recent years. Typically, the existing methods are designed for nonhierarchical labeled data. Such methods fail to exploit the intercategory correlations in the label hierarchy and, therefore, cannot achieve optimal performance on hierarchical labeled data. To address this problem, we propose a novel metric learning method for hierarchical labeled multimodal data, named deep hierarchical multimodal metric learning (DHMML). It learns the multilayer representations for each modality by establishing a layer-specific network corresponding to each layer in the label hierarchy. In particular, a multilayer classification mechanism is introduced to enable the layerwise representations to not only preserve the semantic similarities within each layer, but also retain the intercategory correlations across different layers. In addition, an adversarial learning mechanism is proposed to bridge the cross-modality gap by producing indistinguishable features for different modalities. Through integration of the multilayer classification and adversarial learning mechanisms, DHMML can obtain hierarchical discriminative modality-invariant representations for multimodal data. Experiments on two benchmark datasets are used to demonstrate the superiority of the proposed DHMML method over several state-of-the-art methods.

Sounding Robots: Design and Evaluation of Auditory Displays for Unintentional Human-robot Interaction

Non-verbal communication is important in HRI, particularly when humans and robots do not need to actively engage in a task together, but rather they co-exist in a shared space. Robots might still need to communicate states such as urgency or availability, and where they intend to go, to avoid collisions and disruptions. Sounds could be used to communicate such states and intentions in an intuitive and non-disruptive way. Here, we propose a multi-layer classification system for displaying various robot information simultaneously via sound. We first conceptualise which robot features could be displayed (robot size, speed, availability for interaction, urgency, and directionality); we then map them to a set of audio parameters. The designed sounds were then evaluated in five online studies, where people listened to the sounds and were asked to identify the associated robot features. The sounds were generally understood as intended by participants, especially when they were evaluated one feature at a time, and partially when they were evaluated two features simultaneously. The results of these evaluations suggest that sounds can be successfully used to communicate robot states and intended actions implicitly and intuitively.

A Multi-Layer Classifier Model XR-KS of Human Activity Recognition for the Problem of Similar Human Activity

Sensor-based human activity recognition is now well developed, but there are still many challenges, such as insufficient accuracy in the identification of similar activities. To overcome this issue, we collect data during similar human activities using three-axis acceleration and gyroscope sensors. We developed a model capable of classifying similar activities of human behavior, and the effectiveness and generalization capabilities of this model are evaluated. Based on the standardization and normalization of data, we consider the inherent similarities of human activity behaviors by introducing the multi-layer classifier model. The first layer of the proposed model is a random forest model based on the XGBoost feature selection algorithm. In the second layer of this model, similar human activities are extracted by applying the kernel Fisher discriminant analysis (KFDA) with feature mapping. Then, the support vector machine (SVM) model is applied to classify similar human activities. Our model is experimentally evaluated, and it is also applied to four benchmark datasets: UCI DSA, UCI HAR, WISDM, and IM-WSHA. The experimental results demonstrate that the proposed approach achieves recognition accuracies of 97.69%, 97.92%, 98.12%, and 90.6%, indicating excellent recognition performance. Additionally, we performed K-fold cross-validation on the random forest model and utilized ROC curves for the SVM classifier to assess the model’s generalization ability. The results indicate that our multi-layer classifier model exhibits robust generalization capabilities.

Transformer-based descriptors with fine-grained region supervisions for visual place recognition

Visual place recognition is a fundamental component in autonomous systems and robotics, which is easily limited in the real world with different viewpoints and changes in appearance. Existing approaches to tackle this problem mainly rely on dominant CNN-based architectures, which are difficult to model global correlation. Most recently, there has emerged little work that focuses on the effectiveness of the Transformer in modeling long-range dependencies, but this strategy ignores local interactions thus fails to localize the really important regions. To address the above issue, this paper proposes an effective Transformer-based architecture that takes full advantages of the strengths of Transformer related to global context modeling and local specific region capturing. We first design a dual-level Transformer descriptor encoder to successively perform self-attention within local windows and global extent of the CNN feature map to obtain multi-scale spatial context, which combines local interaction and global information. Specifically, multi-layer classification tokens from the Transformer encoder are integrated to form the global image representation. Moreover, a Transformer-guided geometric verification module is introduced to leverage the strengths of the hierarchical Transformer’s inherent self-attention mechanism for fusing multi-level attention, which is employed to filter the output token to obtain key patches and associated attention weights for achieving spatial matching. Finally, we propose a descriptor refinement strategy that employs fine-grained region-level supervisions to further enhance the capability of the network to learn local discriminative features, which effectively alleviates the confusion caused from weak image-level labels. Extensive experiments on benchmark datasets show that our approach outperforms state-of-the-art methods with promising performance.

CIRCE: Web-Based Platform for the Prediction of Cannabinoid Receptor Ligands Using Explainable Machine Learning.

The endocannabinoid system, which includes cannabinoid receptor 1 and 2 subtypes (CB1R and CB2R, respectively), is responsible for the onset of various pathologies including neurodegeneration, cancer, neuropathic and inflammatory pain, obesity, and inflammatory bowel disease. Given the high similarity of CB1R and CB2R, generating subtype-selective ligands is still an open challenge. In this work, the Cannabinoid Iterative Revaluation for Classification and Explanation (CIRCE) compound prediction platform has been generated based on explainable machine learning to support the design of selective CB1R and CB2R ligands. Multilayer classifiers were combined with Shapley value analysis to facilitate explainable predictions. In test calculations, CIRCE predictions reached ∼80% accuracy and structural features determining ligand predictions were rationalized. CIRCE was designed as a web-based prediction platform that is made freely available as a part of our study.

Deep learning‐based high‐dimensional multiple regression estimator for chest x‐ray image classification in rapid cardiomegaly screening

AbstractChest x‐ray (CXR) examination is a common first‐line, non‐invasive, and rapid screening method in clinical examinations. The posteroanterior (PA) and anteroposterior (AP) view modes can both be used to detect related cardiopulmonary diseases, such as pneumonitis, tuberculosis, pulmonary fibrosis, lung tumors, and cardiomegaly. Compared with cardiac computed tomography and cardiac magnetic resonance imaging methods, CXR examination has a short scanning duration and costs less, and is suitable for routine and follow‐up health examinations. Cardiomegaly is an asymptomatic disease in the early stage and cannot be detected through electrocardiography measurements. Thus, early cardiomegaly classes detections, such as cardiac hypertrophy and ventricular dilatation, can help make decisions regarding drug treatments and surgeries. In addition, an automatic assistive tool is required to differentiate between normal individuals and those with cardiomegaly to address the problem of manual inspection and labor shortage. Hence, PA view‐based CXR classification is used to develop a deep learning (DL)‐based high‐dimensional multiple regression analysis (MRA) model for CXR image classification in rapid cardiomegaly screening. This multilayer network model uses a two‐channel three‐layer convolution‐normalization‐pooling process with two‐dimensional (2D) multi convolution operations to enhance images and to extract feature patterns; and then a one‐dimensional feature conversion is used to estimate the four coordinate points of the maximal horizontal cardiac diameter (MHCD) and maximal horizontal thoracic diameter (MHTD), which can be used to estimate cardiothoracic ratio and detect cardiomegaly. For experimental tests, the training and testing datasets are collected from the National Institutes of Health CXR Image Database (Clinical Center, USA), and 10‐fold cross‐validation was used for model evaluation in terms of precision (%), recall (%), accuracy (%), and F1 score. These indexes are used to evaluate the feasibility of the proposed MRA estimator. In addition, the performances of the proposed model are compared with those of conventional DL‐based multilayer classifiers.

Trusted Fine-Grained Image Classification through Hierarchical Evidence Fusion

Fine-Grained Image Classification (FGIC) aims to classify images into specific subordinate classes of a superclass. Due to insufficient training data and confusing data samples, FGIC may produce uncertain classification results that are untrusted for data applications. In fact, FGIC can be viewed as a hierarchical classification process and the multilayer information facilitates to reduce uncertainty and improve the reliability of FGIC. In this paper, we adopt the evidence theory to measure uncertainty and confidence in hierarchical classification process and propose a trusted FGIC method through fusing multilayer classification evidence. Comparing with the traditional approaches, the trusted FGIC method not only generates accurate classification results but also reduces the uncertainty of fine-grained classification. Specifically, we construct an evidence extractor at each classification layer to extract multilayer (multi-grained) evidence for image classification. To fuse the extracted multi-grained evidence from coarse to fine, we formulate evidence fusion with the Dirichlet hyper probability distribution and thereby hierarchically decompose the evidence of coarse-grained classes into fine-grained classes to enhance the classification performances. The ablation experiments validate that the hierarchical evidence fusion can improve the precision and also reduce the uncertainty of fine-grained classification. The comparison with state-of-the-art FGIC methods shows that our proposed method achieves competitive performances.

Leveraging relevant summarized information and multi-layer classification to generalize the detection of misleading headlines

Disinformation is an important problem facing society nowadays. Given the rapid and easy access to information, news stories quickly go viral, the vast majority of which are misleading and with no prospect of verification. Specifically, the headline of a correctly designed news item must correspond to a summary of the main information of that news item and it should be neutral. However, many headlines circulating on the Internet use false or distorted information, seeking to confuse or mislead the reader. Misleading headlines indicate a dissonance between the headline and the content of the news story. From a computational perspective, this problem is being tackled as a Stance Detection problem between the headline and the body text of the news item. This paper contributes to the fight against the spread of misleading information by presenting a generic and flexible multi-level hierarchical classification. The approach is based on two stages that enable the detection of the stance between the news headline and the body text. The proposed architecture, called HeadlineStanceChecker+ uses the headline and only the essential information of the news item (not the full body text) as inputs. To extract this essential information, different summarization approaches (extractive and abstractive) are analyzed in order to determine the most relevant information for the task. The experimentation has been carried out using the Fake News Challenge (FNC-1) dataset. A 94.49% accuracy was obtained using extractive summaries, which were more helpful than abstractive ones. HeadlineStanceChecker+ improves the accuracy results of existing state-of-the-art systems. In conclusion, using automatic extractive summaries together with the two-stage generic architecture is an effective solution to the problem.

The Application of Machine Learning InSpeech Emotion Recognition

In many applications, such as voice assistants, call centers, psychological counseling, business negotiation, and even hostage rescue, it is becoming more and more important to know the mental state of the other side of the conversation. This article introduces a Speech emotion recognition project based on python and Librosa. This project uses machine learning to train computers to judge the emotional state of a speaker from human speech. Based on this project, the accuracy and efficiency of different models as well as potential application directions are discussed. This project uses an artificial neural network (ANN) model to train the data through a multi-layer classifier (MLP). The ANN model used in this project and the influence of different parameters in the model are discussed, and higher accuracy is obtained on the basis of existing models and data. The article will focus on the analysis of the model structure and the effects of different parameters in the model and their corresponding optimal intervals.

An Adaptive Real-Time Malicious Node Detection Framework Using Machine Learning in Vehicular Ad-Hoc Networks (VANETs)

Modern vehicle communication development is a continuous process in which cutting-edge security systems are required. Security is a main problem in the Vehicular Ad Hoc Network (VANET). Malicious node detection is one of the critical issues found in the VANET environment, with the ability to communicate and enhance the mechanism to enlarge the field. The vehicles are attacked by malicious nodes, especially DDoS attack detection. Several solutions are presented to overcome the issue, but none are solved in a real-time scenario using machine learning. During DDoS attacks, multiple vehicles are used in the attack as a flood on the targeted vehicle, so communication packets are not received, and replies to requests do not correspond in this regard. In this research, we selected the problem of malicious node detection and proposed a real-time malicious node detection system using machine learning. We proposed a distributed multi-layer classifier and evaluated the results using OMNET++ and SUMO with machine learning classification using GBT, LR, MLPC, RF, and SVM models. The group of normal vehicles and attacking vehicles dataset is considered to apply the proposed model. The simulation results effectively enhance the attack classification with an accuracy of 99%. Under LR and SVM, the system achieved 94 and 97%, respectively. The RF and GBT achieved better performance with 98% and 97% accuracy values, respectively. Since we have adopted Amazon Web Services, the network's performance has improved because training and testing time do not increase when we include more nodes in the network.

Detection of Abnormal Network Traffic Using Bidirectional Long Short-Term Memory

Nowadays, web systems and servers are constantly at great risk from cyberattacks. This paper proposes a novel approach to detecting abnormal network traffic using a bidirectional long short-term memory (LSTM) network in combination with the ensemble learning technique. First, the binary classification module was used to detect the current abnormal flow. Then, the abnormal flows were fed into the multilayer classification module to identify the specific type of flow. In this research, a deep learning bidirectional LSTM model, in combination with the convolutional neural network and attention technique, was deployed to identify a specific attack. To solve the real-time intrusion-detecting problem, a stacking ensemble-learning model was deployed to detect abnormal intrusion before being transferred to the attack classification module. The class-weight technique was applied to overcome the data imbalance between the attack layers. The results showed that our approach gained good performance and the F1 accuracy on the CICIDS2017 data set reached 99.97%, which is higher than the results obtained in other research.

Multi-Tier Sentiment Analysis of Social Media Text Using Supervised Machine Learning

Sentiment Analysis (SA) is often referred to as opinion mining. It is defined as the extraction, identification, or characterization of the sentiment from text. Generally, the sentiment of a textual document is classified into binary classes i.e., positive and negative. However, fine-grained classification provides a better insight into the sentiments. The downside is that fine-grained classification is more challenging as compared to binary. On the contrary, performance deteriorates significantly in the case of multi-class classification. In this study, pre-processing techniques and machine learning models for the multi-class classification of sentiments were explored. To augment the performance, a multi-layer classification model has been proposed. Owing to similitude with social media text, the movie reviews dataset has been used for the implementation. Supervised machine learning models namely Decision Tree, Support Vector Machine, and Naïve Bayes models have been implemented for the task of sentiment classification. We have compared the models of single-layer architecture with multi-tier model. The results of Multi-tier model have slight improvement over the single-layer architecture. Moreover, multi-tier models have better recall which allow our proposed model to learn more context. We have discussed certain shortcomings of the model that will help researchers to design multi-tier models with more contextual information.

Evaluating the Effectiveness of Interactive Video Learning by Examining Machine Learning Classifiers Models: Graduate Students’ Perspectives

Important elements had an impact on how traditional learning was implemented and motivated researchers to develop Interactive Video Learning Effectiveness (IVL-E). These variables range from price to learning-environment to learner perspective, among others. This paper’s major objectives are to: (i) assess the effectiveness of Interactive Video Learning (IVL-E) using classification techniques and considering graduate students’ viewpoints, (ii) establish appropriate classification parameters to choose the optimum classifier model, and (iii) review prior works pertaining to IVL-E assessment. The study dataset is a sample of 63 datapoints randomly chosen by a survey performed at the College of Education, University of Bisha, Bisha, Saudi Arabia. A total of 123 registered postgraduate students made up the study population when using Google’s online questionnaire method, after all the respondents voluntarily agreed to fill out and submit the questionnaire. This study develops a reliable machine learning classifier’s model for classifying IVL-E. The created models use a backpropagation algorithm and are a type of multilayer classification perceptron. The best classification output was “interactive video learning performance measure”, which provided the highest results under: 1) support vector machine-based classifier (SVC), 2) decision tree (DT), and 3) light gradient-boosting machine classifier (lgb.LGBMClassifier). Regarding classification measures like balanced accuracy (high BCCR = 0.875), balanced error rate (low BER = 0.125), and optimization precision (highest OP = 0.999), our models performed extremely well according to the literature review.

Automatic Segmentation of Calcified Plaques and Vessel Borders from Abdominal Aorta CTA Images Using a 2D Spatial Convolutional Multilayer Machine Vision Classifier

The aorta, including the thoracic and abdominal aortas, distributes blood from the heart to the upper and lower circulatory systems. Abdominal aortic sclerosis can cause abdominal aortic aneurysm. When the diameter of the aorta increases by more than 1.5 times its normal size (e.g., ≥4−5 cm, mean diameter in an adult is 2−3 cm), the risk of aneurysm rupture gradually increases. Aneurysm rupture and its associated bleeding have high mortality rates. Hence, the prediction of large diameter aortic aneurysms is an important issue with great clinical implications. In computed tomography angiography (CTA) imaging examination, an assistive tool for automatic image segmentation must be developed to help cardiologists and radiologists evaluate the risk of rupture progression and make appropriate treatment decisions. In this study, two‐dimensional (2D) spatial convolution operations (corner detection) and a multilayer machine vision classifier enabled the automatic segmentation of the luminal region, vascular intima, and calcified plaques in the abdominal aorta. The 2D spatial convolution operations can efficiently increase focus, remove unwanted noise in the convolutional layer, and segment calcified plaques and the vessel border by using a general regression neural network (GRNN)‐based classifier. The classifier's outputs can also be visualized in color presentation (red, green, and blue color series) to identify vessel borders and calcified plaques. Luminal change rate (LCR), calcified plaque rate (CPR), and ellipse ratio (ER) can be defined as criteria for evaluating changes in the luminal region and locating calcified plaques, which can be used to identify the locations of cross‐sectional shape changes and calcified plaques. Based on the collected CTA images from the clinical case studies, the proposed multilayer classifier has promising results for the intended medical purpose. With the LCR index, the correlation coefficients (R2) for manual image segmentation and automatic image segmentation is >0.90 in positive correlation for vessel borders segmentation. The proposed method is an adaptive learning scheme with potential clinical applications in automatic CTA image segmentation for further abdominal aortic aneurysm (AAA) risk evaluation. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

SiamMBFAN: Siamese tracker with multi-branch feature aggregation network

Siamese trackers have attracted considerable attention in the field of object tracking because of their high precision and speed. However, one of the main disadvantages of Siamese trackers is that their feature extraction network is relatively single. They often use AlexNet or ResNet50 as the backbone network. AlexNet is shallow and thus cannot easily extract abundant semantic information, whereas ResNet50 has many convolutional layers, reducing the real-time performance of Siamese trackers. We propose a multi-branch feature aggregation network with different designs in the shallow and deep convolutional layers. We use the residual module to build the shallow convolutional layers to extract textural and edge features. The deep convolution layers, designed with two independent branches, are built with residual and parallel modules to extract different semantic features. The proposed network has a depth of only nine modules, and thus it is a simple and effective network. We then apply the network to a Siamese tracker to form SiamMBFAN. We design multi-layer classification and regression subnetworks in the Siamese tracker by aggregating the last three modules of the two branches, improving the localization ability of the tracker. Our tracker achieves a better balance between performance and speed. Finally, SiamMBFAN is tested on four challenging benchmarks, including OTB100, VOT2016, VOT2018, and UAV123. Compared with other trackers, our tracker improves by 7% (OTB100).

Assessment of urban air quality from Twitter communication using self-attention network and a multilayer classification model.

Social media platforms are one of the prominent new-age methods used by public for spreading awareness or drawing attention on an issue or concern. This study demonstrates how the twitter responses of public can be used for qualitative monitoring of air pollution in an urban area. Tweets discussing about air quality in Delhi, India, were extracted during 2019–2020 using a machine learning technique based on self-attention network. These tweets were cleaned, sorted, and classified into 3-class quality viz. poor air quality, good air quality, and noise or neutral tweets. The present study used a multilayer classification model with first layer as an embedding layer and second layer as bi-directional long-short term memory (BiLSTM) layer. A method was then devised for estimating PM2.5 concentration from the tweets using ‘spaCy’ similarity analysis of classified tweets and data extracted from Continuous Ambient Air Quality Monitoring Stations (CAAQMS) in Delhi for the study period. The accuracy of this estimation was found to be high (80–99%) for extreme air quality conditions (extremely good or severe) and lower during moderate variations in air quality. Application of this methodology depended on perceivable changes in air quality, twitter engagement, and environmental consciousness among public.Supplementary InformationThe online version contains supplementary material available at 10.1007/s11356-022-22836-w.

Time-Series Monitoring of Dust-Proof Nets Covering Urban Construction Waste by Multispectral Images in Zhengzhou, China

Rapid urbanization has produced a huge amount of construction waste. The operations and consequences of construction and demolition can lead to windblown dust problems, profoundly affecting the living environment of residents. Fortunately, dust-proof nets have been used in construction sites to reduce and prevent pollution by fine particles such as dust, so it is important to monitor and evaluate their efficacy. In this study, Earth observation techniques were used for the extraction and monitoring of solid waste and dust-proof nets. In order to fully perceive the validity and necessity of dust-proof nets for urban air health, we conducted a case study in Zhengzhou, China. We explored the potential of multispectral remote sensing available for monitoring urban construction waste and proposed a multi-layer classification method to identify construction waste and dust-proof nets based on Landsat-8 OLI and Sentinel-2 MSI data, with an average identification accuracy and Kappa coefficient of 96.27% and 0.93 for construction waste in the study area from 2015 to 2020, respectively. In addition, our study revealed the driving factors and impact of temporal variations in regional construction waste areas and dust-proof nets coverage. The results indicate the classification can track municipal solid waste management and changes in air pollutant concentrations and is useful for achieving SDG 11.6 (reduce the adverse per capita environmental impact of cities, including by paying special attention to air quality and municipal and other waste management). This study has the potential to monitor construction waste and dust-proof nets, paving the way for better urban environmental governance and surveillance actions in the future, especially involving big data.

Fuzzy Sematic Segmentation and Efficient Classification of Lung Cancer Multi-Dimensional Datasets

Lung cancer is one of the leading cause of cancer death around the world. Lung cancer has been the most common cancer worldwide since 1985, both in terms of incidence and mortality. Recognition and prediction of lung cancer at the earliest stage can be very useful to improve the survival rate of patients. Effective and early diagnosis of cancer is one the major challenging task for medical practitioners. In this research work, we propose a novel technique on lung MRI image based segmentation and classification is using fuzzy logic and deep learning. The proposed technique considers multi-dimensional medical dataset modeling and representation for effective diagnosis and prediction. A fuzzy based sematic segmentation with relevance to Region of Interest (RoI) extraction and append deep learning models to customized RoI selection under segmented patches. The multi-layer classification approach is viewed to be an effective and accurate diagnosis method for the prediction of disease at early stage.

Multivariate multi-layer classifier

The variance-ratio binary multi-layer classifier (VRBMLC) has been recently proposed and shown to outperform conventional binary decision trees (BDTs). Though effective with better interpretability, the VRBMLC generates deep layers of tree nodes as it employs a one-feature-at-a-time binary split at each layer. To further condense the tree depth and enhance the classification performance, this research proposes a multivariate multi-layer classifier that applies a variance-ratio criterion to enable ternary splits of each tree node and that integrates the oblique discriminant hyperplane in the tree node. We benchmark 16 state-of-the-art univariate and multivariate classifiers on 43 publicly available datasets. The results show that the proposed methods greatly simplify the tree structure and yield a significantly higher average accuracy.

Sparse data classifier based on the first-past-the-post voting system

Point of Interest (POI) is a general term for objects describing places from the real world. The concept of POIs matching, i.e. determining whether two sets of attributes represent the same location, is not a trivial challenge due to the large variety of data sources. The representation of POIs may vary depending on the base in which they are stored. Manual comparison of objects with each other is not achievable in real-time, therefore there are multiple solutions to automatic merging. However there is no efficient solution that includes the deficiencies in the existence of attributes, has been proposed so far. In this paper, we propose the Multilayered Hybrid Classifier which is composed of machine learning and deep learning techniques, supported by the first-past-the-post voting system. We examined different weights for constituencies which were taken into consideration during the majority (or supermajority) decision. As a result, we achieved slightly higher accuracy than the current best model - Random Forest, which in its working also base on voting.