Network Learning Research Articles

Learning personalized drug features and differentiated drug-pair interaction information for drug–drug interaction prediction

Multi-drug combination therapies are increasingly used for complex diseases but carry risks of harmful drug interactions. Effective drug–drug interaction prediction (DDIP) is essential for assessing risks among numerous drug pairs. Most DDIP methods involve two main steps: drug representation and drug pair interaction extraction, respectively challenged by the loss of personalized drug information and the need for differentiated interaction data, and are rarely studied. Specifically, personalized drug information refers to the distinct features of each drug. These properties can be easily confused by neighboring information during graph propagation. This issue is especially prominent in drug interaction graphs with long-tail distributions, which poses challenges for personalized drug information learning. Furthermore, it is crucial to learn interactions with differentiation in order to identify diverse drug relationships. Some methods simply concatenate drug features, often ignoring the differences of different drug relationships, while other methods based on substructures rely on professional pharmacological knowledge and are computationally complex. To address these issues, we propose a novel method, learning personalized Drug Features and differentiated Drug-Pair interaction information for drug–drug interaction prediction (DFPDDI). This approach employs a contrastive learning network with edge-aware augmentations and mutual information estimators to capture personalized drug features across various graph distributions. Furthermore, it applies a mutual information constraint to drug-pair representations, enhancing the accuracy of interaction predictions by better distinguishing between different types of drug relationships. The results evaluated on three public datasets demonstrate competitive performance compared to baselines. It also shows potential for accurate predictions, particularly in imbalanced-distribution graphs.

A lightweight network-based sign language robot with facial mirroring and speech system

Human–robot interaction is an essential capability for humanoid robots to enter the physical world and become companions in people’s lives, learning, and work. While the majority of current research focuses on the voice-based interactions of robots, yet over 60% of communication occurs through nonverbal behaviors, such as facial expressions and hand gestures. Endowing robots with the ability to communicate through nonverbal behavior not only enhances the interactive experience with robots but also provides a potential communication tool for individuals with hearing or speech impairments. Here, we develop a humanoid robot capable of adjusting facial movements by driving servos, and design a novel framework for the robot to integrate sign language recognition and facial landmark detection algorithms. This framework facilitates the robot recognize sign language and translate it into spoken language, while also imitating the facial expressions of the signers. To achieve this, we also propose a lightweight deep learning network called RealTimeSignNet for real-time sign language recognition. Leveraging lightweight 3D convolution modules and time-dependent constraints, this model adapts to various time scales, ensuring efficient processing of sign language recognition tasks. Experimental results demonstrate the outstanding performance of the RealTimeSignNet model on mainstream sign language datasets, achieving an accuracy of 88.1% on the large continuous sign language dataset (continuous SLR), 98.2% on the isolated sign language dataset (SLR 500), and 91.50% on the English sign language dataset (WLAS). The overall assessment demonstrates that our humanoid robot is capable of recognizing sign language and translating it into spoken language, while imitating the facial emotions, providing a comprehensive solution to the communication challenges faced by individuals with hearing and speech impairments.

Rapid Response to Professional Development: How Online Teacher Professional Learning Networks Supported Teachers in Pandemic Times

During COVID-19, when teachers rapidly shifted to remote teaching and learning, it became evident that there was an educational crisis—teachers were unprepared for teaching using digital pedagogies. A group of teachers and teacher educators in Trinidad and Tobago, ITTPN Global, responded to the call for emergency professional development in ICT tools with a free workshop series via Zoom. The workshop series served to address teachers’ upskilling in key areas of ICT as digital pedagogy became the pandemic pedagogy. Pandemic pedagogy refers to teaching and learning during COVID-19. Teachers built confidence and competence in using ICT tools and designing online lessons. The professional learning model highlighted in this paper utilized authentic and active learning and teacher agency. This reflective essay examines the transformative possibilities of networks in a crisis with implications for professional teacher learning post pandemic.

Improving the UNet with channel attention mechanism for Taihu Lake water body recognition

Abstract Accurate recognition of surface water bodies is important for disaster management and resource planning. To explore a higher-precision water body recognition method, propose an improved channel attention mechanism UNet (ECA-UNet) water body recognition model based on Sentinel-2A remote sensing images. Utilizing ECA-UNet, UNet, PSPnet, and DeeplabV3, as well as the traditional normalized water body index (NDWI), recognize the water body in the Taihu Lake area. Through comparative analysis, the following main conclusions are drawn: (1) The recognition accuracy of the deep learning network model is better than the traditional NDWI method, among which the ECA-UNet network model has the highest accuracy of 98.67%, and the lowest recognition accuracy of NDWI is 89.06%. (2) The ECA-UNet network model can effectively extract water body recognition targets, has a better recognition effect on local features of water bodies, and can improve the water body extraction effect. The above results show that the improved channel attention mechanism ECA-UNet has the highest accuracy in water body recognition in Taihu Lake, which provides a reference for subsequent research and applications in this field.

A Lightweight Recurrent Learning Network for Sustainable Compressed Sensing

Recently, deep learning-based compressed sensing (CS) has achieved great success in reducing the sampling and computational cost of sensing systems and improving the reconstruction quality. These approaches, however, largely overlook the issue of the computational cost; they rely on complex structures and task-specific operator designs, resulting in extensive storage and high energy consumption in CS imaging systems. In this article, we propose a lightweight but effective deep neural network based on recurrent learning to achieve a sustainable CS system; it requires a smaller number of parameters but obtains high-quality reconstructions. Specifically, our proposed network consists of an initial reconstruction sub-network and a residual reconstruction sub-network. While the initial reconstruction sub-network has a hierarchical structure to progressively recover the image, reducing the number of parameters, the residual reconstruction sub-network facilitates recurrent residual feature extraction via recurrent learning to perform both feature fusion and deep reconstructions across different scales. In addition, we also demonstrate that, after the initial reconstruction, feature maps with reduced sizes are sufficient to recover the residual information, and thus we achieved a significant reduction in the amount of memory required. Extensive experiments illustrate that our proposed model can achieve a better reconstruction quality than existing state-of-the-art CS algorithms, and it also has a smaller number of network parameters than these algorithms. Our source codes are available at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/C66YU/CSRN</uri> .

SCANet: A lightweight deep learning network for massive MIMO CSI feedback based on spatial and channel attention mechanism

Multi-user massive multiple-input multiple-output (MIMO) communication systems consume too much downlink bandwidth due to the huge channel state information (CSI) feedback, deep learning-based CSI feedback approaches fortunately can alleviate the feedback overhead while obtaining an accurate CSI. However, there is a trade-off between the high feedback performance and low computational complexity. In this paper, a low-complexity CSI feedback approach is proposed based on spatial and channel attention mechanism, namely the Spatial and Channel Attention Network (SCANet). Specifically, the spatial and channel attention mechanism makes the network’s attention mainly focus on the specific spatial regions and key feature channels. We devise a serial architecture in the encoder that composes of Spatial and Channel Attention Block (SCAB) and Encoder Transformer Block. Moreover, we design a hybrid architecture in the decoder that composes of the CNNs Block and Decoder Transformer Block. These designs enable the network to effectively extract both global and local CSI features. Computer simulations in both the indoor and outdoor scenarios show that under the same system configurations, the proposed low-complexity SCANet achieves almost the same performance as the state-of-the-art network while reducing the computational complexity by 85.76% fewer floating-point operations per second (FLOPS) on average.

Mpox outbreak: Time series analysis with multifractal and deep learning network.

This article presents an overview of an mpox epidemiological situation in the most affected regions-Africa, Americas, and Europe-tailoring fractal interpolation for pre-processing the mpox cases. This keen analysis has highlighted the irregular and fractal patterns in the trend of mpox transmission. During the current scenario of public health emergency of international concern due to an mpox outbreak, an additional significance of this article is the interpretation of mpox spread in light of multifractality. The self-similar measure, namely, the multifractal measure, is utilized to explore the heterogeneity in the mpox cases. Moreover, a bidirectional long-short term memory neural network has been employed to forecast the future mpox spread to alert the outbreak as it seems to be a silent symptom for global epidemic.

A multi-label legal charge prediction method based on summarization generation

Legal Charge Prediction (LCP) is a task of predicting charge labels for which a criminal should be charged based on the factual case descriptions. In recent years, the advancement of deep learning networks has led to the emergence of effective approaches for LCP through the Fine-tuning of pre-trained language models. While existing methods primarily treat it as multiple binary classifications, excelling in the single charge label given factual case descriptions, they still exhibit substantial room for improvement in multi-label LCP. This is because single-label classification methods do not effecitively deal with the relationship between multiple labels, mainly determining the likelihood of a factual description’s association to each individual charge label. Recognizing the need to enhance the relationship among multi-label charges in Fine-tuning-based discriminative charge prediction, this paper introduces a multi-label LCP method based on summarization generation. This method reconstructs the multi-label classification problem as a summarization generation task, capturing the relationship among multiple labels within the summarization content. Empirical findings from the Chinese legal domain’s CAIL2018 dataset demonstrate the method’s capacity to discern the relationship between multiple legal charge labels with Micro-F1 and Macro-F1 enhancements, compared to the fine-tuning-based approaches.

Multi-scale fusion pixel and instance contrastive self-supervised learning for semantic segmentation of high-resolution Earth surface images

In the field of remote sensing (RS) image semantic segmentation, existing supervised learning methods rely on a large amount of labeled data, which limits their application scope. To solve this problem, we propose a new self-supervised learning method, called multi-scale fusion pixel and instance contrastive learning network (MPINet). This method first uses focal frequency loss to optimize the learning of high-level semantic information, and then strengthens the spatial information in the shallow feature map through multi-scale fusion pixel contrast learning, thereby improving the model’s ability to mine detailed features. Experiments are conducted on the International Society for Photogrammetry and Remote Sensing (ISPRS) Potsdam, LoveDA and UAVid datasets. The results show that our method achieves 49.89%, 70.98% and 63.55% in mIoU, OA and mF1 indicators on the ISPRS Potsdam dataset, which is 1.48%, 1.71% and 0.95% higher than the best method. On the LoveDA dataset, our method achieves 38.05%, 52.89% and 54.10% in mIoU, OA and mF1 indicators, which is 0.6%, 0.91% and 1.04% higher than the best method. On the UAVid dataset, our method achieved mIoU, OA, and mF1 scores of 57.92%, 76.97%, and 73.28%, respectively, representing improvements of 1.06%, 1.46%, and 1.15% compared to the best method. Experimental results show that the proposed method outperforms the existing optimal self-supervised learning method and the ImageNet pre-training method.

A deep convolutional neural network for Blind element Error Correction of Spatial Heterodyne Spectrometer using line selective convolutional blocks

The "GF Special Project" is a massive remote sensing technology initiative including a number of satellites and various observation platforms. GF-5 is the satellite with the most payloads, the highest spectral resolution, and the most difficulty in development, and it can monitor a variety of environmental elements using spatial heterodyne spectroscopy (SHS) technology, including atmospheric aerosols, carbon dioxide, methane, terrestrial vegetation, straw burning, and urban heat islands. In this study, a novel blind element error correction technique based on deep learning network is investigated and developed for spatial heterodyne interferograms, as well as the formation mechanism and distribution characteristics of the SHS interferometric data. LSConv-Net, a new CNN model, was created and trained to denoise in the presence of high-density and ultra-high-density blind element errors. We do this by introducing a new line-selective convolutional (LSConv) block. Simultaneously, experimental validation of blind element error correction utilizing laboratory water vapor interferometric data and atmospheric CO2 absorption interferometric data from GF-5, and the change in FWHM before and after the experiment was tested using potassium lamp interferograms. Experiments show that the Deep neural networks trained with this model may successfully suppress the effect of blind element noise on spectra, recover spectra that have been overwhelmed by high-density blind element noise without any effect on other non-blind pixels, and surpass all similar techniques in terms of spectral recovery.

Mutitask Learning Network for Partial Discharge Condition Assessment in Gas-Insulated Switchgear

Mutitask Learning Network for Partial Discharge Condition Assessment in Gas-Insulated Switchgear

Leak detection in water supply pipeline with small-size leakage using deep learning networks

Pipeline transportation technology has witnessed remarkable development in recent years, while the issue of leakage poses a serious limitation to this technology’s progress, especially for the small-scale leakage in water pipelines that exhibits weaker leakage signals. To address those challenges, this study investigates leak detection in water supply pipeline with an inner pipe diameter of 100 mm in the operating range of volume flow Q = 40–80 m3/h, leakage flow Ql = 0.65–1.33 m3/h and pressure p = 130–220 kPa. The characteristics and propagation of leakage pressure signals are analyzed. An adaptive improved variational mode decomposition (VMD) based on the white shark optimizer (WSO) algorithm is proposed, utilizing sample entropy as the fitness function. Additionally, tent chaotic mapping is employed for parameter initialization, mean normalization is used to minimize the influence of the DC component, and signal relative energy serves as a boundary condition. The processed data is utilized to establish leakage detection models using long short-term memory network (LSTM), gated recurrent unit (GRU), convolutional neural network (CNN) and SqueezeNet framework. The performance of leakage detection models is evaluated and compared using 962 sets’ data using evaluation indicators and confusion matrix. The results indicate that SqueezeNet-GRU demonstrates superior performance in pipeline leak detection and localization applications, achieving 97.40 % accuracy for the total dataset.

Transportation Mode Recognition Based on Low-Rate Acceleration and Location Signals With an Attention-Based Multiple-Instance Learning Network

Transportation Mode Recognition Based on Low-Rate Acceleration and Location Signals With an Attention-Based Multiple-Instance Learning Network

IRDNet: An image-defogging algorithm based on dark channel prior

Abstract Image-defogging, as an important part of computer vision, has been widely used in intelligent driving, target recognition, satellite detection, underwater exploration and so on. Improving the performance of the defogging algorithm based on deep learning has practical significance for the completion of high-level vision tasks. This paper proposes an IRDNet algorithm by improving the image deep learning algorithm based on the RefineDNet framework and designing a new deep learning network structure. The proposed algorithm combines dark channel prior knowledge and atmospheric degradation model to decompose the input fogged images into dark channel images and degraded images, and then carries out feature extraction and detail enhancement for the two images by convolutional neural network. IRDNet introduces a fully connected convolutional structure, attention mechanism, and a pyramid structure to improve the performance of the overall defogging network. By testing on OTS datasets, we compare the IRDNet algorithm with traditional defogging methods, end-to-end deep learning defogging methods, and prior knowledge-guided deep learning defogging methods. Test results show that the performance of IRDNet is better than other methods, and the defogging network shows richer details of texture and more realistic colors.

Similarity learning networks uniquely identify individuals of four marine and terrestrial species

AbstractEstimating the size of animal populations plays an important role in evidence‐based conservation and management. Some methods for estimating population size rely on animals being individually identifiable. Traditionally, this has been done by marking physically captured animals, but increasingly, animals with distinctive natural markings are surveyed noninvasively using cameras. Animal reidentification from photographs is usually done manually, which is expensive, laborious, and requires considerable skill. An alternative is to develop computer vision methods that can support or replace the manual identification task. We developed an automated approach using deep learning to identify whether a pair of photographs is of the same individual or not. The core of the approach is a similarity learning network that uses paired convolutional neural networks with a triplet loss function to summarize image pairs and decide whether they are from the same individual. Prior to the main matching step, two additional convolutional neural networks perform image segmentation, cropping the animal object within the image, and orientation prediction, deciding which side of the animal was photographed. We applied the approach to four species, with images of the same individual often spanning several years: systematic surveys of bottlenose dolphins (Tursiops truncatus, 2008–2019) and harbor seals (Phoca vitulina, 2015–2019), a citizen science dataset of western leopard toads (Sclerophrys pantherina, unknown dates), and a publicly available repository of humpback whale images (Megaptera novaeangliae, unknown dates). For these species, our best‐performing models were able to identify whether a pair of images were from the same individual or different individuals in 95.8%, 94.6%, 88.2%, and 83.8% of the cases, respectively. We found that triplet loss functions outperformed binary cross‐entropy loss functions and that data augmentation and additional manual curation of training data provided small but consistent improvements in performance. These results demonstrate the potential of deep learning to replace or, more likely, support and facilitate manual individual identification efforts.

Proton dose calculation on cone-beam computed tomography using unsupervised 3D deep learning networks

Background and PurposePoor image quality of cone-beam computed tomography (CBCT) images can hinder proton dose calculation to assess the influence of anatomy changes. The aim of this study was to evaluate image quality and proton dose calculation accuracy of synthetic CTs generated from CBCT using unsupervised 3D deep-learning networks. Materials and methodsA total of 102 head-and-neck cancer patients were used to train (N=82) and test (N=20) i) a cycle-consistent generative adversarial network, ii) a contrastive unpaired translation, and iii) a fusion of the two (CycleCUT). For patients in the test set, a repeat CT was deformably registered to a same-day CBCT to create a ground-truth CT for comparison. The proton plan was re-calculated on the ground-truth CT and synthetic CTs. The image quality of the synthetic CTs was evaluated using peak signal-to-noise ratio, structural similarity index measure, mean error, and mean absolute error (MAE). Proton dose calculation accuracy was assessed through 3D gamma analysis and dose-volume-histogram parameters. ResultsAll synthetic CTs accurately preserved the CBCT anatomy (verified by visual inspection) while improving the image quality. The CycleCUT network had slightly improved image quality compared to the other networks (MAE in body: 53 Hounsfield units (HU) vs. 54/55 HU). All networks had similar proton dose calculation accuracy with gamma passing rate above 97%. ConclusionsAll three evaluated networks generated synthetic CT images with dose distributions comparable to those of conventional fan-beam CT. The synthetic CT generation was fast, making all networks feasible for adaptive proton therapy.

RAF-Unet: A Remote Sensing Identification Method for Forest Land Information with Modified Unet

Abstract Carrying out remote sensing refinement identification of forest land in complex environment is of great significance for timely mapping of forest distribution. Aiming at the problem that remote sensing images have bias in the extraction of forest land information data, based on the semantic segmentation algorithm Unet, combining the ResNet50 deep learning network, the attention mechanism module and the feature pyramid structure, we construct RAF-Unet (ResNet+Attention+FPN+Unet) to improve the extraction of forest land information data. The ResNet50 classification network is used as the encoder of the Unet network to extract the feature maps at five different scales; then, the attention mechanism module is introduced in the decoder stage of the Unet network to extract the key task goal information by learning the weight values of the features; finally, the feature pyramid structure is used in the output stage of the encoder to fuse the information from the shallow network and the deep network to extract the remote sensing forest land information in the image. The results show that the RAF-Unet algorithm outperforms the Unet algorithm in all the indexes, with a precision of 95.24%, a recall of 91.80%, an F1-score value of 93.49%, an intersection over union of 87.63%, and an accuracy of 93.68%; the validity of the modules is verified by the ablation experiments, and the ResNet network, the attention mechanism, and the feature pyramid structure are all effective in improve the classification effect. It helps the forestry department to better manage and dynamically monitor forestry information, which is of great significance to the scientific development, utilization and protection of forest land resources.

Modified local Granger causality analysis based on Peter‐Clark algorithm for multivariate time series prediction on IoT data

AbstractClimate data collected through Internet of Things (IoT) devices often contain high‐dimensional, nonlinear, and auto‐correlated characteristics, and general causality analysis methods obtain quantitative causality analysis results between variables based on conditional independence tests or Granger causality, and so forth. However, it is difficult to capture dynamic properties between variables of temporal distribution, which can obtain information that cannot be obtained by the mean detection method. Therefore, this paper proposed a new causality analysis method based on Peter‐Clark (PC) algorithm and modified local Granger causality (MLGC) analysis method, called PC‐MLGC, to reveal the causal relationships between variables and explore the dynamic properties on temporal distribution. First, the PC algorithm is applied to compute the relevant variables of each variable. Then, the results obtained in the previous stage are fed into the modified local Granger causality analysis model to explore causalities between variables. Finally, combined with the quantitative causality analysis results, the dynamic characteristic curves between variables can be obtained, and the accuracy of the causal relationship between variables can be further verified. The effectiveness of the proposed method is further demonstrated by comparing it with standard Granger causality analysis and a two‐stage causal network learning method on one benchmark dataset and two real‐world datasets.

Prediction of tool wear and surface finish using ANFIS modelling during turning of Carbon Fiber Reinforced Plastic (CFRP) composites

Carbon fiber-reinforced plastics (CFRP) are widely used in various industries due to their high strength to weight ratio, corrosion resistance, durability, and excellent thermo-mechanical properties. The machining of CFRP composites has always been a challenge for the manufacturers. In this study, CNC turning operation with coated carbide tool is used to machine a specific CFRP and the relationship between the cutting parameters (Speed, Feed, Depth of Cut) and response parameters (Vibration, Surface Finish, Cutting Force and Tool Wear) are investigated. An adaptive-network-based fuzzy inference system (ANFIS) model with two multi-input–single-output (MISO) system has been developed to predict the tool wear and surface finish. Speed, feed, depth of cut, vibration and cutting force have been used as input parameters and tool wear and surface finish have been used as output parameters. Three sets of cutting parameter have been used to gather the data points for continuous turning of CFRP composite. The model merged fuzzy inference modeling with artificial neural network learning abilities, and a set of rules is constructed directly from experimental data. This model is capable of predicting the cutting tool wear and surface finish during turning of CFRP composite. The predicted tool wear and surface finish data are compared to the experimental results. The predicted data agreed well with the actual experimental data with 98.96 % accuracy for tool wear and 99.61 % accuracy for surface finish.

Hierarchical Organ-Aware Total-Body Standard-Dose PET Reconstruction From Low-Dose PET and CT Images.

Positron emission tomography (PET) is an important functional imaging technology in early disease diagnosis. Generally, the gamma ray emitted by standard-dose tracer inevitably increases the exposure risk to patients. To reduce dosage, a lower dose tracer is often used and injected into patients. However, this often leads to low-quality PET images. In this article, we propose a learning-based method to reconstruct total-body standard-dose PET (SPET) images from low-dose PET (LPET) images and corresponding total-body computed tomography (CT) images. Different from previous works focusing only on a certain part of human body, our framework can hierarchically reconstruct total-body SPET images, considering varying shapes and intensity distributions of different body parts. Specifically, we first use one global total-body network to coarsely reconstruct total-body SPET images. Then, four local networks are designed to finely reconstruct head-neck, thorax, abdomen-pelvic, and leg parts of human body. Moreover, to enhance each local network learning for the respective local body part, we design an organ-aware network with a residual organ-aware dynamic convolution (RO-DC) module by dynamically adapting organ masks as additional inputs. Extensive experiments on 65 samples collected from uEXPLORER PET/CT system demonstrate that our hierarchical framework can consistently improve the performance of all body parts, especially for total-body PET images with PSNR of 30.6 dB, outperforming the state-of-the-art methods in SPET image reconstruction.