The small satellite market continues to grow year after year. A compound annual growth rate of 17% is estimated during the period between 2020 and 2025. Low-cost satellites can send a vast amount of images to be post-processed at the ground to improve the quality and extract detailed information. In this domain lies the resolution enhancement task, where a low-resolution image is converted to a higher resolution automatically. Deep learning approaches to Super-Resolution (SR) reached the state-of-the-art in multiple benchmarks; however, most of them were studied in a single-frame fashion. With satellite imagery, multi-frame images can be obtained at different conditions giving the possibility to add more information per image and improve the final analysis. In this context, we developed and applied to the PROBA-V dataset of multi-frame satellite images a model that recently topped the European Space Agency’s Multi-frame Super Resolution (MFSR) competition. The model is based on proven methods that worked on 2D images tweaked to work on 3D: the Wide Activation Super Resolution (WDSR) family. We show that with a simple 3D CNN residual architecture with WDSR blocks and a frame permutation technique as data augmentation better scores can be achieved than with more complex models. Moreover, the model requires few hardware resources, both for training and evaluation, so it can be applied directly from a personal laptop.

The COVID-19 pandemic situation has created even more difficulties in the quick identification and screening of the COVID-19 patients for the medical specialists. Therefore, a significant study is necessary for detecting COVID-19 cases using an automated diagnosis method, which can aid in controlling the spreading of the virus. In this paper, the study suggests a Deep Convolutional Neural Network-based multi-classification approach (COV-MCNet) using eight different pre-trained architectures such as VGG16, VGG19, ResNet50V2, DenseNet201, InceptionV3, MobileNet, InceptionResNetV2, Xception which are trained and tested on the X-ray images of COVID-19, Normal, Viral Pneumonia, and Bacterial Pneumonia. The results from 3-class (Normal vs. COVID-19 vs. Viral Pneumonia) showed that only the ResNet50V2 model provides the highest classification performance (accuracy: 95.83%, precision: 96.12%, recall: 96.11%, F1-score: 96.11%, specificity: 97.84%) compared to rest of the models. The results from 4-class (Normal vs. COVID-19 vs. Viral Pneumonia vs. Bacterial Pneumonia) demonstrated that the pre-trained model DenseNet201 provides the highest classification performance (accuracy: 92.54%, precision: 93.05%, recall: 92.81%, F1-score: 92.83%, specificity: 97.47%). Notably, the ResNet50V2 (3-class) and DenseNet201 (4-class) models in the proposed COV-MCNet framework showed higher accuracy compared to the rest six models. This indicates that the designed system can produce promising results to detect the COVID-19 cases on the availability of more data. The proposed multi-classification network (COV-MCNet) significantly speeds up the existing radiology-based method, which will be helpful to the medical community and clinical specialists for early diagnosis of the COVID-19 cases during this pandemic.

During the period January 2014 – October 2018, four strong earthquakes occurred in the Ionian Sea, Greece. A rich aftershock sequence followed each event of them. More analytically, according to the manual solutions of National Observatory of Athens, the first event (K1), occurred on 26 January 2014 in Kefallinia Island with magnitude M_L = 5.8, which was followed by another in the same region (K2) on 3 February 2014 with magnitude M_L = 5.7. The third event occurred on 17 November 2015, M_L = 6.0 in Lefkas Island (L1) and the last on 25 October 2018, M_L = 6.6 in Zande Island (Z1). The first three of these earthquakes caused moderate structural damages mainly in houses and produced particular unrest to the local population. This work presents first the calculation of the source parameters of the strong events as well as for all earthquakes with magnitude M_L > 4.0, using the methodology of the Moment tensor inversion and secondary data from permanent GPS stations were analyzed to confirm the findings from seismological data and to investigate the displacement due to the earthquakes.

Cyber security plays an important role to protect our computer, network, program and data from unauthorized access. Intrusion detection system (IDS) and intrusion prevention system (IPS) are two main categories of cyber security, designed to identify any suspicious activities present in inbound and outbound network packets and restrict the suspicious incident. Deep neural network plays a significant role in the construction of IDS and IPS. This paper highlights a novel IDS using optimized convolution neural network (CNN-IDS). An optimized CNNIDS model is an improvement over CNN which selects the best weighted model by considering the loss in every epoch. All the experiments have been conducted on the well known NSL-KDD dataset. Information gain has been used for dimensionality reduction. The accuracy of the proposed model is evaluated through optimized CNN for both binary and multiclass categories. Finally, a critical comparison has been performed with other general classifiers like J48, Naive Bayes, NB tree, Random forest, Multilayer Perceptron (MLP), Support Vector Machine (SVM), Recurrent Neural Network (RNN) and Convolution Neural Network(CNN). All the experimental results demonstrate that the optimized CNN-IDS model records the best recognition rate with minimum model construction time.

For enhancing the maximized profit from bank as well as customer perspective, term deposit can accelerate finance fields. This paper focuses on likelihood of term deposit subscription taken by the customers. Bank campaign efforts and customer details are influential while considering possibilities of taking term deposit subscription. An automated system is provided in this paper that approaches towards prediction of term deposit investment possibilities in advance. Neural network(NN) along with stratified 10-fold cross-validation methodology is proposed as predictive model which is later compared with other benchmark classifiers such as k-Nearest Neighbor (k-NN), Decision tree classifier (DT), and Multi-layer perceptron classifier (MLP). Experimental study concluded that proposed model provides significant prediction results over other baseline models with an accuracy of 88.32% and Mean Squared Error (MSE) of 0.1168.

This article presents an automated vision-based algorithm for the die-scale inspection of wafer images captured using scanning acoustic tomography (SAT). This algorithm can find defective and abnormal die-scale patterns, and produce a wafer map to visualize the distribution of defects and anomalies on the wafer. The main procedures include standard template extraction, die detection through template matching, pattern candidate prediction through clustering, and pattern classification through deep learning. To conduct the template matching, we first introduce a two-step method to obtain a standard template from the original SAT image. Subsequently, a majority of the die patterns are detected through template matching. Thereafter, the columns and rows arranged from the detected dies are predicted using a clustering method; thus, an initial wafer map is produced. This map is composed of detected die patterns and predicted pattern candidates. In the final phase of the proposed algorithm, we implement a deep learning-based model to determine defective and abnormal patterns in the wafer map. The experimental results verified the effectiveness and efficiency of our proposed algorithm. In conclusion, the proposed method performs well in identifying defective and abnormal die patterns, and produces a wafer map that presents important information for solving wafer fabrication issues.

With the development of information technology and the popularization of mobile devices, collecting various types of customer data such as purchase history or behavior patterns became possible. As the customer data being accumulated, there is a growing demand for personalized recommendation services that provide customized services to customers. Currently, global e-commerce companies offer personalized recommendation services to gain a sustainable competitive advantage. However, previous research on recommendation systems has consistently raised the issue that the accuracy of recommendation algorithms does not necessarily lead to the satisfaction of recommended service users. It also claims that customers are highly satisfied when the recommendation system recommends diverse items to them. In this study, we want to identify the factors that determine customer satisfaction when using the recommendation system which provides personalized services. To this end, we developed a recommendation system based on Deep Neural Networks (DNN) and measured the accuracy of recommendation service, the diversity of recommended items and customer satisfaction with the recommendation service. The experimental results of is the study showed that both recommendation system accuracy and diversity would have a positive effect on customer satisfaction. These results can further improve customer satisfaction with the recommendation system and promote the sustainable development of e-commerce.

Timely detection of surface damages on wind turbine blades is imperative for minimising downtime and avoiding possible catastrophic structural failures. With recent advances in drone technology, a large number of high-resolution images of wind turbines are routinely acquired and subsequently analysed by experts to identify imminent damages. Automated analysis of these inspection images with the help of machine learning algorithms can reduce the inspection cost, thereby reducing the overall maintenance cost arising from the manual labour involved. In this work, we develop a deep learning based automated damage suggestion system for subsequent analysis of drone inspection images. Experimental results demonstrate that the proposed approach could achieve almost human level precision in terms of suggested damage location and types on wind turbine blades. We further demonstrate that for relatively small training sets advanced data augmentation during deep learning training can better generalise the trained model providing a significant gain in precision.

This study was undertaken to forecast waste generation rates of accommodation sector of North Cyprus as a case. Three predictor models including multiple linear regression (MLR), artificial neural networks (ANNs) and central composite design (CCD) were applied to predict the waste generation rate during the lean and peak seasons. ANN showed highest prediction performance, specifically, lowest values of the standard error of prediction (SEP = 2.153), mean absolute error (MAE=1.378) and highest R2 value (0.998) confirmed the accuracy of the model. The analysed wastes were categorised into recyclable, general waste and food residues. The authors estimated the total waste generated during the lean season as 2010.5 kg/day, in which large-sized hotel accounted for largest fraction (66.7%), followed by the medium hotels (19.4%) and guesthouse accounted for smallest part (2.6%). During the peak season, about 49.6% increases in the waste generation rates were obtained. Interestingly, 45% of the wastes were generated by the British tourists while the least waste was generated by African tourists (7.5%). The ANN predicted that the small and large hotels would produce 5.45 and 22.24 tons of waste by the year 2020, respectively. The findings herein are promising and useful in establishing a sustainable waste management system.

Deep learning has recently attracted much attention due to its excellent performance in processing audio, image, and video data. However, few studies are devoted to the field of automatic modulation classification (AMC). It is one of the most well-known research topics in communication signal recognition, which remains challenging for traditional methods due to the complex disturbance from other sources. This paper proposes a heterogeneous deep model fusion (HDMF) method to solve the problem in a unified framework. The contributions include: 1) The convolutional neural network (CNN) and long short-term memory (LSTM) are combined by two different ways without prior knowledge involved; 2) A large database, including eleven types of single-carrier modulation signals with various noises as well as a fading channel, is collected with various signal-to-noise ratios (SNRs) based on a real geographical environment; and 3) Experimental results demonstrate that HDMF is super capable of copping with the AMC problem, and achieves much better performance when compared with the independent network. The source code and the database will be publically available.

As AI models become more and more common in business and even in our daily lives, it is important to understand what the carbon impact of these models is. Recent papers have shown that this impact can be quite great, i.e., the training of a single high-end model can result in emissions of more than 500t of CO2eq. In this paper we discuss the factors that influence the carbon footprint of AI models, explore what impact different decisions have, and show how the footprint can be reduced. We also examine the footprint of different models to give a guideline on how urgent action is for different organizations.

To address the challenge of recognizing ship states accurately amidst the complexities of marine environments, this study proposes a novel ship state recognition approach leveraging a graph convolutional neural network (GCNN). Initially, the method extracts canonical and efficient ship motion trajectories from AIS data. Subsequently, a state recognition network tailored for ship motion trajectories is devised and implemented employing graph convolution. Notably, the accuracy of this model is enhanced through the introduction of novel weights and optimization of the Adj parameter. Experimental evaluations conducted on a ship state dataset demonstrate significant performance improvements. Specifically, the proposed recognition network achieves a recognition accuracy of 98.3% for regulated ship trajectories, marking an impressive 8.4% enhancement over traditional convolutional neural networks. This advancement holds promise for enhancing ship state recognition accuracy across diverse maritime applications including maritime supervision, navigation safety, and ship management.

Mechanical equipment is composed of several parts, the interaction between the parts exists throughout whole life cycle leads to the widespread phenomenon of fault coupling and the diagnosis of independent faults cannot meet the requirements of the health management of mechanical equipment under actual working conditions. In this paper, the topological structure of graph neural network is used to describe the correlation of coupling faults and multiple fault types are regarded as network nodes respectively. The fault features are fused and extracted by graph convolution to achieve the classification of coupling faults. In this paper, dynamic vertexes are defined in data topology and the parameters of DIGNN classification model are optimized by changing the fault type of dynamic vertexes during training stage. The data is loaded into dynamic vertexes for classification and analysis during testing stage. The one-dimensional vibration data is converted into two-dimensional time-frequency domain data by wavelet transform making the input features of the model interpretable to reduce the uncertainty of model training. The data topology is then fed into DIGNN for fault classification, and GCN's node aggregation gives interpretability to each layer of network data processing. The method proposed in this paper can realize accurate diagnosis of independent faults on the data set, and can effectively judge the coupling mode of coupling faults, which is suitable for coupling fault analysis of mechanical equipment.

Imitation Learning is a discipline of Machine Learning primarily concerned with replicating observed behavior of agents known to perform well on a given task, collected in demonstration data sets. In this paper, we set out to introduce a pipeline for collecting demonstrations and training models that can produce motion plans for industrial robots. Object throwing is defined as the motivating use case. Multiple input data modalities are surveyed, and motion capture is selected as the most practicable. Two model architectures operating autoregressively are examined -- feedforward and recurrent neural networks. Trained models execute throws on a real robot successfully, and a battery of quantitative evaluation metrics is proposed, including extrapolated throw accuracy estimates. Recurrent neural networks outperform feedforward ones in most respects, with the best models having an assessed mean throw error on the order of 0.1...0.2 m at distances of 1.5...2.0 m. The data collection, pre-processing, and model training aspects of our proposed approach show promise, but further work is required in developing Cartesian motion planning tools before it is suitable for application in production.

Traffic shaping is a critical task in software-defined -IoT networks (SDN-IoTs) to efficiently manage network resources and ensure Quality of Service (QoS) for end-users. However, traditional traffic shaping approaches based on queuing theory or static policies may not be effective due to the dynamic and unpredictable nature of network traffic. In this paper, we propose a novel approach that leverages Graph Neural Networks (GNNs) and Multi-arm Bandit algorithms to dynamically optimize traffic shaping policies based on real-time network traffic patterns. Specifically, our approach uses a GNN model to learn and predict network traffic patterns and a Multi-arm Bandit algorithm to optimize traffic shaping policies based on these predictions. We evaluate the proposed approach on three different datasets, including a simulated corporate network (KDD Cup 1999), a collection of network traffic traces (CAIDA), and a simulated network environment with both normal and malicious traffic (NSL-KDD). The results demonstrate that our approach outperforms other state-of-the-art traffic shaping methods, achieving higher throughput, lower packet loss, and lower delay, while effectively detecting anomalous traffic patterns. The proposed approach offers a promising solution to traffic shaping in SDNs, enabling efficient resource management and QoS assurance

Stemming from the object overlap and undertraining from the few samples, the road dense object detection is confronted with the poor object identification performance and the inability to recognize edge objects. Based on this, one transfer learning-based YOLOv3 approach for identifying dense objects in the road has been proposed. Firstly, Darknet-53 network structure is adopted to obtain pre-trained YOLOv3 model, then the transfer training is introduced as the output layer for the special dataset of 2000 images containing vehicles; in the proposed model, one random function is adapted to intialize and optimize the weights of the transfer training model, which is seperately designed from the pre-trained YOLOv3; and the object detection classifier replaces the fully connected layer, which further improves the detection effect, the reduced size of the network model can further reduce the training and detection time, and can be better applied to actual scenarios. The experimental results demonstrate that the object detection accuracy of the presented approach is 87.75% for the Pascal VOC 2007 dataset, which is superior to the traditional YOLOv3 and the traditional YOLOv2 by 3.05% and 11.15%, respectively. Besides, the test was carried out using UA-DETRAC, a public road vehicle detection dataset, the object detection accuracy of the presented approach reaches 79.23% in detecting images, which is 4.13% better than the traditional YOLOv3, and compared with the existing relatively new object detection algorithm YOLOv5, the detection accuracy is 1.36% better. Moreover, the detection speed of the proposed YOLOv3 method reaches 31.2 Fps/s in detecting images, which is 7.6 Fps/s faster than the traditional YOLOv3, and compared with the existing relatively new object detection algorithm YOLOv5, the speed is 4.3 Fps/s faster; the proposed YOLOv3 performs 79.38Bn of floating point operations per second in detecting video, which obviously surpasses the traditional YOLOv3 and the newer object detection algorithm YOLOv5.

We introduce a new, and quite general variational model for opinion dynamics based on pairwise interaction potentials and a range of opinion evolution protocols ranging from random interactions to global synchronous flows in the opinion state space. The model supports the concept of topic “coupling”, allowing opinions held by individuals to be changed via indirect interaction with others on different subjects. Interaction topology is governed by a graph that determines interactions. Our model, which is really a family of variational models, has, as special cases, many of the previously established models for the opinion dynamics. After introducing the model, we study the dynamics of the special case in which the potential is either a tent function or a constructed bell-like curve. We find that even in these relatively simple potential function examples there emerges interesting behavior. We also present results of preliminary numerical explorations of the behavior of the model to motivate questions that can be explored analytically

This paper analyses the application of deep learning techniques for predicting wave overtopping events in port environments using sea state and weather forecasts as inputs. The study was conducted in the outer port of Punta Langosteira, A Coruña, Spain. A video recording infrastructure was installed to monitor overtopping events from 2015 to 2022, identifying 3709 overtopping events. The data collected was merged with actual and predicted data for the sea state and weather conditions during the overtopping events, creating three datasets. We used these datasets to create several machine learning models to predict whether an overtopping event would occur based on sea state and weather conditions. The final models achieved a high accuracy level during the training and testing stages: 0.81, 0.73, and 0.84 average accuracy during training and 0.67, 0.48, and 0.86 average accuracy during testing, respectively. The results of this study have significant implications for port safety and efficiency, as wave overtopping events can cause disruptions and potential damage. Using deep learning techniques for overtopping prediction can help port managers take preventative measures and optimize operations, ultimately improving safety and helping to minimize the economic impact that overtopping events have on the port's activities.

Object detection is a popular image processing technique, widely used in numerous applications for detecting and locating objects in images or videos. While being one of the fastest algorithms for object detection, Single-Shot Multibox Detection (SSD) networks are also computationally very demanding, which limits their usage in real-time edge applications. Even though the SSD post-processing algorithm is not the most complex segment of the overall SSD object detection network, it is still computationally demanding and can become a bottleneck with respect to processing latency and power consumption, especially in edge applications with limited resources. When using hardware accelerators to accelerate backbone CNN processing, the SSD post-processing step implemented in software can become critical for high-end applications where high frame rates are required, as this paper shows. To overcome this problem, we propose Puppis, an architecture for hardware acceleration of SSD post-processing algorithm. As experiments will show, our solution will lead to an average SSD post-processing speedup of 34.42 when compared with a software implementation. Furthermore, execution of a complete SSD network will be on average 45.39 times faster than software implementation when the proposed Puppis SSD hardware accelerator is used together with some existing CNN accelerators.

Identifying influential actors within social networks is pivotal for optimizing information flow and mitigating the spread of both rumors and diseases. Several methods have emerged to pinpoint these influential entities in networks that are represented as graphs. In these graphs, nodes corre-spond to individuals and edges indicate their connections. This study focuses on centrality measures, prized for their straightforwardness and effectiveness. We categorize structural central-ity into two: local, considering a node's immediate vicinity, and global, accounting for overarching path structures. Some techniques blend both centralities to highlight nodes influential at both mi-cro and macro levels. Our paper presents a novel centrality measure, accentuating node degree and incorporating the network's broader features, especially paths of different lengths. Through Spearman and Pearson's correlations tested on seven standard datasets, our method proves its merit against traditional centrality measures. Additionally, we employ the SIR model, portraying disease spread, to further validate our approach. The ultimate influential node is gauged by its capacity to infect the most nodes during the SIR model's progression. Our results indicate a nota-ble correlative efficacy across various real-world networks relative to other centrality metrics.

Retweet prediction is an important task related to different problems such as information spreading analysis, the automatic detection of fake news, social media monitoring, etc. In this study we explore the possibilities of retweet prediction based on heterogeneous data sources. In order to classify the tweet according to the amount of retweets, we combine features extracted from the multilayer network and the text. More specifically, we introduce a multilayer framework that proposes the multilayer network representation of Twitter. This formalism captures different users' actions and complex relationships as well as other key properties of communication on Twitter. We select a set of local network measures from each layer and construct a set of multilayer network features. In addition, we adopt a BERT-based language model, namely Cro-CoV-cseBERT to capture high-level semantics and structure of tweets as a set of text features. Then, we train six machine learning (ML) algorithms: random forest, multilayer perceptron, light gradient boosting machine, category embedding model, neural oblivious decision ensembles and attentive interpretable tabular learning model in the task of retweet prediction. We compare the performance of all six algorithms in three different setups (i) using only text features, (ii) using only multilayer network features and (iii) using both sets of features. We evaluate all setups in terms of standard evaluation measures i.e. precision, recall, F1-score and accuracy. For this task, we first prepare and use an empirical dataset of 199,431 tweets in the Croatian language posted during the period between January 1, 2020 and May 31, 2021. Our results indicate that by integrating multilayer network features with text features the prediction model would perform better than using just one set of features.

Abstract: To describe the feeder topology and information exchange between intelligent electronic devices (IEDs) for Internet of Things (IoT), it is necessary to extend IEC 61850 and add a logical model. The feeder model described in IEC 61970 is then extended in IEC 61850 to serve as a container for the equipment along the distribution line. For the physical connections in the electrical equipment, both terminal and connectivity node logical models are added. Taking logical node XCBR as the example, its attributes for the terminals are added to the electrical equipment. The logical node LCNN is then further added in order to describe the connectivity node. By next adding the logical node LTPN to describe the topological node, and FTPA to describe the results of the topology analysis, we can efficiently describe the topology analysis results. Based on these new logical nodes, the exchange of topology information between intelligent terminals no longer relies on configuration files, but only requires logical nodes. Through the above expansion, the topology description and information exchange of medium and low voltage feeders are realized.

Snow parameters have traditionally been retrieved using discontinuous, multi-band sensors; however, continuous hyperspectral sensor are now being developed as an alternative. In this paper we investigate the performance of various sensor configurations using machine learning neural networks trained on a simulated dataset. Our results show improvements in accuracy of retrievals of snow grain size and impurity concentration for continuous hyperspectral channel configurations. Retrieval accuracy of snow albedo was found to be similar for all channel configurations.

With the widespread use of new equipment such as distributed photovoltaics, distributed energy storage, electric vehicles, and distributed wind power, the control of low-voltage distribution networks (LVDNs) has become increasingly complex. Acquiring the most recent topological structure is essential for conducting accurate analysis and real-time control of LVDNs. The signal-injection-based topology identification algorithm is favored for its speed and efficiency. This study proposes a new signal-injection-based topology identification algorithm aimed at solving the issues of reduced completeness and correctness of identification caused by missing feature signal records (FSRs). By analyzing the correlations between FSRs of different positions, the algorithm employs vertical and horizontal completion methods to effectively address missing data, combined with inclusion detection algorithms, to identify the topology of LVDNs. Based on the study of actual LVDN data, the results indicate that the algorithm not only significantly improves the completeness and correctness of topology identification for LVDNs but can also increase the number of devices in operation by evaluating the status of topological equipment, further enhancing the topology identification performance.

Recently, inter-vehicle communication, which helps to avoid collision accidents (by driving safety support system) and facilitate self-driving (by dissemination of road and traffic information), has attracted much attention. In this paper, in order to efficiently collect road / traffic information in the request / response manner, first a basic method, Content-centric network (CCN) for Vehicular network (CV), is proposed, which applies CCN cache function to inter-vehicle communication. Content naming and routing, which take vehicle mobility into account, are investigated. On this basis, the CV method is extended (named as ECV) to avoid the cache miss problem caused by vehicle movement, and is further enhanced (named as ECV+) to more efficiently exploit cache buffer in vehicles, caching content according to a probability decided by channel usage rate. Extensive evaluations on network simulator Scenargie, with realistic open street map, confirm that the CV method and its extensions (ECV, ECV+) effectively reduce the average number of hops of data packets (by up to 47%, 63%, 83% respectively) and greatly improve the content acquisition success rate (by up to 356%, 444%, 689%, respectively), compared to the method without cache mechanism.

Security has become a critical issue in today’s highly distributed and networked systems. Network intrusion detection systems (NIDSs), especially signature-based NIDSs, are being widely deployed in a distributed network environment with the purpose of defending against a variety of network attacks. Most of the commercially available NIDSs are software based and rely on pattern matching to extract the threat from network traffic. The increase in network speed and traffic may make existing algorithms to become a performance bottleneck. Therefore it is very necessary to develop faster and more efficient pattern matching algorithm in order to overcome the troubles on performance of NIDSs. Therefore, we propose a multi fusion pattern matching algorithm for Network Intrusion Detection Systems. The results obtained in percentages from the proposed fusion algorithm given better values in terms processing time in milliseconds than the existing algorithms when data English text are applied to evaluate the fusion performances.

Time series forecasting has been a challenging area in the field of Artificial Intel-ligence. Various approaches such as linear Neural Networks, Recurrent Linear Neural Networks, Convolutional Neural Networks and recently Transformers have been attempted for the time series forecasting domain. Although, Trans-former-based architectures have been outstanding in the Natural Language Pro-cessing domain, especially in autoregressive language modeling, the initial at-tempts to use transformers in the time series arena have met with mixed success. One of the recent papers demonstrated that modeling the multi-variate time se-ries problem as independent channels on a transformer architecture produced lower mean square and mean absolute errors on benchmarks, however, another recent paper claimed that a simple linear neural network out-performs previous approaches including transformer-based designs. We investigate this paradox in detail comparing the linear Neural Network and Transformer based designs, provide insights to why a certain approach may be better for a particular type of problem. We also improve upon the recently proposed simple linear Neural Network based architecture by using dual pipelines with batch normalization and reversible instance normalization. Our enhanced architecture outperforms all existing architectures for time series forecasting on majority of the popular benchmarks.

With the rise of piano teaching in recent years, many people have joined the team of piano learners. However, the expensive cost of manual instruction and the unique one-on-one teaching model have made piano learning an extravagant event. Most existing approaches based on the audio modality aim to evaluate piano players' skills. Unfortunately, these methods ignored the information contained in the video, which led to a one-sided and simplistic evaluation of the piano player's skills. More recently, multimodal-based methods are proposed to assess the skill level of piano players using both video and audio information. However, existing multimodal approaches use shallow networks to extract video and audio features, which are deficient in extracting complex spatio-temporal and time-frequency features from piano performance. Furthermore, the fingering and the pitch-rhythm information of the piano performance is contained in the spatio-temporal and time-frequency features, respectively. In the paper, we propose a ResNet-based audio-visual fusion model that combines video and audio features to assess the skill level of piano players. Firstly, ResNet18-3D is used as the backbone network for our visual branches, which can extract feature information from the video data. Then, we consider ResNet18-2D as the backbone network of the aural branch and extract the feature information from the audio data. The extracted video features are fused with the audio features to generate multimodal features for the final piano skill evaluation. The experimental results on the PISA dataset show that our proposed audio-visual fusion model, with a validation accuracy of 70.80%, outperforms the state-of-the-art methods in both performance and efficiency. Then, we also explore the impact of different layers of ResNet on model performance, and the experimental results show that the audio-visual fusion model dealing with the piano skill assessment problem can make full use of both feature information when the number of video features is close to the number of audio features.

Retinopathy is a human eye disease that causes changes in retinal blood vessels that leads to bleed, leak fluid and vision impairment. Symptoms of retinopathy are blurred vision, changes in color perception, red spots, and eye pain. In this paper, a new methodology based on Convolutional Neural Networks (CNN) is developed and proposed to diagnose and give a decision about the presence of retinopathy. The CNN model is trained by different images of eyes that have retinopathy and those which do not have retinopathy. The performance of the proposed model is compared with the related methods of DREAM, KNN, GD-CNN and SVM. Experimental results show that the proposed CNN performs better.

Abstract: With the growing integration of the Internet of Things (IoT) in smart buildings, it is crucial to ensure the precise implementation and operation of wireless sensor networks (WSNs). This paper aims to study the implementation aspect of Wireless Sensor Networks (WSNs) in a commercial multi-story building, specifically addressing the difficulty of dealing with variable environmental conditions on each floor. This research addresses the disparity between simulated situations and actual deployments, offering valuable insights that have the potential to signifi-cantly improve the efficiency and responsiveness of building management systems. We obtain real-time sensor data to analyze and evaluate the system’s performance. Our investigation is grounded in the growing importance of incorporating WSNs into buildings to create intelligent environments. By leveraging the capabilities of Contiki operating system, this study employs a specialized simulator to generate simulated data, allowing for a comparison between real-world and simulated scenarios. We provide an in-depth analysis to scrutinize the disparities and commonalities between the datasets obtained from real-world deployments and simulated en-vironments. We highlight the necessity of precise simulation to validate real-life sensor data, shedding light on the challenges and opportunities associated with implementing WSNs in in-telligent buildings. Results obtained show the significance of accurate simulation models for re-liable data representation, providing a roadmap for further developments in the integration of WSNs into intelligent building scenarios. The findings highlight the potential for optimizing living and working conditions based on the real-time monitoring of critical environmental pa-rameters. This includes insights into temperature, humidity, and light intensity, offering oppor-tunities for enhanced comfort and efficiency in intelligent environments.

Embedding representation models characterize each word as a vector of numbers with a fixed length. These models have been used in tasks involving text classification, such as recommen- dation and question-answer systems. Semantic relationships are words with a relationship between them providing a complete idea to a text. Therefore, it is hypothesized that an embedding model involving semantic relationships will provide better performance for tasks that use them. This paper presents three embedding models based on semantic relations extracted fromWikipedia to classify texts. The synonym, hyponym, and hyperonym semantic relationships were the ones considered in this work since previous experiments have shown that they are the ones that provide the most semantic knowledge. Lexical-syntactic patterns present in the literature were implemented and subsequently applied to the Wikipedia corpus to obtain the semantic relationships present in it. Several semantic relationships are used in different models: synonymy, hyponym-hyperonym, and a combination of the first two. A convolutional neural network was trained for text classification to evaluate the performance of each model. The results obtained were evaluated with the metrics of precision, accuracy, recall, and F1-measure. The best values obtained with the second model were accuracy of 0.79 for the 20-Newsgroup corpus. F1-measure and recall of 0.87 respectively for the Reuters corpus.

A novel approach is developed to improve the convergence of Physics-Informed Neural Networks (PINNs), aiming to employ them as real-time computational models within the framework of the digital twin for manufacturing processes. This method entails the weighting of physical equations, boundary conditions, and initial conditions to ensure their comparable magnitudes, with power being the chosen quantity in this study. The approach is applied to thermal problems, which are crucial for predicting manufacturing part defects. Different configurations, including complex boundary conditions and complex physics, were tested to assess the model’s robustness. The W-PINN demonstrates good predictions and strong stability compared to the classical PINN.

Accurate forecasting of available energy portion that corresponds to the reservoir inflow of the month(s) ahead provides important decision support for the hydropower plants in energy pro-duction planning for revenue maximization, as well as for the environmental impact prevention, and flood control at the upstream and downstream of a basin. Therefore, a reliable forecasting tool or model is deemed necessary and crucial. Considering the fluctuation and nonlinearity of data which significantly influence the forecasting results, this study develops an effective hybrid model by integrating Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO) called “PSO-ANN” model based on the hydrological and meteorological data pre-processed by cross-correlation function (CCF), autocorrelation function (AFC), and normalization techniques for predicting the available energy portion corresponding to the reservoir inflow mentioned above for a case study hydropower plant in Laos namely, Theun-Hinboun hydropower plant (THHP). The model was evaluated by using correlation coefficient (r), relative error (RE), root mean square error (RMSE), and Taylor diagram plots in comparison with the popular single algorithm approaches such as ANN, and NARX models. Results demonstrated the superiority of the proposed PSO-ANN approach over the other two models.

Early disease detection is crucial for maximizingcrop yields and minimizing financial losses in agriculture. Inorder to categorise leaf diseases in pepper, tomato, and potato plants, this study suggests a revolutionary CNN design. The algorithm collects disease-specific features from input photos by using data augmentation techniques to artificially enlarge the training dataset. ReLU-activated convolutional layers gradually capture features at different levels, from low to high, while max pooling layers minimise spatial dimensionality. After thecharacteristics are extracted, a fully-connected layers with a softmax activation function classify them into illness groups. A well chosen dataset containing labelled healthy and sick leaves is used to train and validate the model. Generalizability to untesteddata is ensured via performance evaluation on an independent testing set. The field of deep learning for agricultural applications benefits from this research. This strategy has the ability to rev- olutionise agricultural practices by enabling automated disease identification. This might result in early interventions, better crop health, and eventually sustainable agricultural production.

The accurate state of health (SOH) estimation of lithium-ion batteries (LIBs) during operation is crucial to ensure optimal performance, prolonging battery life and preventing unexpected failure or safety hazards. This work presents a storage and performance optimised deep learning approach to predict the capacity based SOH of LIBs using raw sensor data from partial charging curves under constant current condition. The proposed model is based on a combination of a 1-dimensional convolutional and long-short term memory neural network and processes time, voltage and incremental capacity of partial charging curves as time series. The model is cross-validated on different aging scenarios reaching an overall MAE = 0.418% and RMSE = 0.531%, promising an accurate SOH estimation of LIBs under varying usage and environmental conditions in a real world application.

Machine learning (ML) is enabling augmented reality (AR) to gain popularity in various fields, including gaming, entertainment, healthcare, and education. ML enhances AR applications in education by providing accurate visualizations of objects. For AR systems, ML algorithms facilitate the recognition of objects and gestures from kindergarten through university. The purpose of this survey is to provide an overview of various ways in which ML techniques can be applied within the field of AR within education. The first step is to describe the background of AR. In the next step, we will discuss the ML models that are used in AR education applications. Additionally, we discuss how ML is used in AR. Each subgroup’s challenges and solutions can be identified by analyzing these frameworks. In addition, we outline several research gaps and future research directions in ML-based AR frameworks for education.

This paper proposes training improvement methods of artificial neural networks (ANN) trajectory predictors. First, a dynamic power system time-series trajectory is split into several different segments to simplify the original ANN training problem. Moreover, the time-derivative of the trajectory is included to obtain an augmented loss function. Compared to previous studies which mainly focused on increasing the prediction accuracy, the aim of these novel techniques is to reduce the computational burden where the ANN output performance is still acceptable. The effectiveness of the developed methods is validated based on the WSCC three-machine nine-bus and IEEE 39-bus system models. The mean absolute error (MAE) and trajectory prediction results are analyzed, in which the numbers of neurons, hidden layers, and training epochs are constrained during the ANN training process. Rotor-angle difference between generators and the system frequency are investigated as the dynamic trajectories of the power system models. It is revealed that when the ANN architecture and epochs are constrained, the MAE results can be reduced, and the ANN training results can better reflect the original trajectory using the improvement approaches.

Amid the swift progression of the global automobile industry, the aspiration for lightweight vehicle design has surfaced as a crucial approach to boost fuel efficiency and curtail environmental impact. The objective of this study revolves around achieving holistic optimization of vehicle structure via detailed scrutiny of vehicle process characteristics, amalgamated with a Back Propagation neural network (BPNN) optimized by a genetic algorithm (GA) and a topology optimization (TO) method.The process initiates with an examination of the vehicle's technical characteristics, succeeded by an optimization protocol that couples GA with BPNN. This exhaustive optimization procedure derives the best-fit values of design parameters and identifies the globally optimal lightweight design approach. Subsequently, a comprehensive structural optimization is actualized by modifying the topological distribution of the structure.A simulation experiment of the conceptual car body beam skeleton is conducted as a conclusive step, and the anti-collision capacity of the car body beam skeleton is assessed by measuring deformation intrusion. The findings reveal that in the vehicle topology optimization, a comparative analysis of pre and post-optimization results indicates an improved topological density across diverse sections. The topological density of the vehicle's frontal structure escalates from 0.75 to 0.8. Following optimization, the stiffness assessment records a marginal increase (within the range of 0.02–0.05), implying that the lightweight design does not compromise the general rigidity of the structure.The optimized body structure manifests superior crashworthiness in the crash test: the deformation intrusion diminishes by 27% in front-end collisions. In side impacts, the deformation and intrusion at critical points are reduced by 13.15% to 10% respectively. The comprehensive method proposed herein proves to be an efficient strategy to actualize the lightweight design of vehicles.

The integration of Artificial Intelligence (AI) with Digital Twins (DTs) has emerged as a promising approach to revolutionize healthcare, particularly in the diagnosis and management of thoracic disorders. This study proposes a comprehensive framework, named Lung-DT, which leverages IoT sensors and AI algorithms to establish a digital representation of a patient’s respiratory health. Using the YOLOv8 neural network, the Lung-DT system accurately classifies chest X-Rays into five distinct categories of lung diseases, including "Normal," "Covid," "Lung Opacity," "Pneumonia," and "Tuberculosis". The system’s performance was evaluated on a chest X-Ray dataset, demonstrating an impressive average accuracy of 96.6% across all classes. Further tests (prediction) were conducted on the trained network using a third dataset available in the literature and completely unknown to the network, yielding an average accuracy of 98% across three classes. The proposed Lung-DT framework offers several advantages over conventional diagnostic methods. Firstly, it enables real-time monitoring of lung health through continuous data acquisition from IoT sensors, facilitating early diagnosis and intervention. Secondly, the AI-powered classification module provides automated and objective assessments of chest X-Rays, reducing dependence on subjective human interpretation. Thirdly, the twin digital representation of the patient’s respiratory health allows for comprehensive analysis and correlation of multiple data streams, providing valuable insights for personalized treatment plans. The integration of IoT sensors, AI algorithms, and DT technology within the Lung-DT system demonstrates a significant step towards improving thoracic healthcare. By enabling continuous monitoring, automated diagnosis, and comprehensive data analysis, the Lung-DT framework has enormous potential to enhance patient outcomes, reduce healthcare costs, and optimize resource allocation.

This paper focused on the finite element analysis of structural system in extreme loading condition. Two different stud shape and thicknesses were analyzed under blast. The stud thickness such as 1.19 mm and 1.5 mm were modelled and analyzed using ABAQUS 6.14. FEM is a tool which predicts the engineering physics of the real structure. To validate the finite element modelling performed by authors, a reference work published by earlier researchers on cold-formed steel stud wall is considered and examined in the present study. The novelty of this study was web corrugation and influence of flange width on stud. The models mimic like an air bag in a car to delay the pressure timing inside the stud wall. The mass of explosive used as 1.56 kg at a standard scaled distance. Time versus displacement was captured out at four locations in the stud wall. One of the objectives is to develop mathematical model to validate the deformation of stud under blast loading. Two mathematical models were validated using Artificial Neural Network (ANN). The results captured in ANN model was error histogram, regression plot, best performance fit and training data. The models were capable of resisting the moderate blast load. The response surface methodology (RSM) was employed to evaluate model performance Regression equations are useful for predicting future trends and outcomes, which is crucial for planning and decision-making. The primary goal of this work is to evaluate cold-formed steel stud walls with varying stud sizes under blast loading using finite element analysis and validated by ANN and RSM.

This paper presents a new method of process parameters optimization, adequate for 3D printing of PLA (Polylactic Acid) components. The authors developed a new Hybrid Manufacturing Equipment (HME), suitable for producing complex parts made from a biodegradable thermoplastic polymer, for environmental sustainability. Our new HME equipment is producing PLA parts by both additive and subtractive techniques, with the aim of obtaining accurate PLA components with a good surface quality. A design of experiments has been applied for optimization purposes. The following manufacturing parameters were analyzed: rotation of the spindle, cutting depth, feed rate, layer thickness, nozzle speed, and surface roughness. Linear regression models and neural network models were developed to improve and predict the surface roughness of the manufactured parts. A new test part was designed and manufactured from PLA, to validate the new mathematical models, which can now be applied for producing complex parts made from polymer materials.

There is a need for cost-effective and non-invasive methods of monitoring feeding behavior in livestock operations, considering the significant impact of feed costs on economic efficiency and assisting in detecting health issues of group-fed animals. This paper proposes using deep learning-based computer vision techniques to detect pen-fed beef cattle feeding behavior using Mask Region-based Convolutional Neural Network (RCNN). A deep learning model was pre-trained on the Common Objects in Context (COCO) dataset to generate cattle instance segmentation. Manually defined feed bunk polygons are compared with these segmentation masks to derive feeding time for each bunk. A full day’s worth of video data and the corresponding physical sensor data are collected for the experiment. By benchmarking the computer vision detected data with physical ground truth over random time segments from morning to evening (thus various lighting conditions), the optimal thresholds for Mask RCNN are determined to be 0.7 for bounding boxes and 0.1 for masks. Using these parameters. The reports suggest that the computer vision system achieved a precision of 87.2% and a recall of 89.1%, signifying precise detection of feeding events. Our study, to the best of our knowledge, was one of the first investigations of instance segmentation on feeding time sense, which combines deep learning methods with traditional computer vision logistics, reporting on feeding time data collection and processing, camera testing and adjustment, and performance evaluation. Future research directions include computer vision applied in feed grading and animal re-identification for individual production analysis.

For Convolutional Neural Network (CNN), Depthwise Separable CNN (DSCNN) is a preferred architecture for Application Specific Integrated Circuit (ASIC) implementation on edge devices. It can benefit from a multi-mode approximate multiplier proposed in this work. The proposed approximate multiplier uses two 4-bit multiplication operations to implement a 12-bit multiplication operation by reusing the same multiplier array. With this approximate multiplier, sequential multiplication operations are pipelined in a modified DSCNN to fully utilize the PE array in the convolutional layer. This Approximate (A-DSCNN) was implemented on TSMC 40-nm CMOS process with a supply voltage of 0.9 V. At the clock frequency of 200 MHz, the design achieves 4.78 GOPs/mW while occupying 1.24 mm x 1.24 mm silicon area. Compared to conventional DSCNN implemented in a similar process node, the chip area and power consumption were reduced by 53% and 25%, while the throughput was improved by 17%.

In a limestone quarry mine, fragmentation is a crucial outcome of blasting operations. The optimization of blasting operations greatly benefits from the prediction of rock fragmentation. The main factors that affect fragmentation are rock mass characteristics, blast geometry, and explosive properties. This paper is a step towards the implementation of machine learning and deep learning algorithms for predicting the extent of fragmentation (in percentage) in opencast mining. Various parameters can affect fragmentation. But, in this paper initially, ten parameters (spacing, drill hole diameter, burden, average bench height, powder factor, number of holes, charge per delay, uniaxial compressive strength, specific drilling, and stemming) are collected to train the model. However, due to a weak correlation with rock fragmentation, drill diameter, Average bench height, compressive strength, stemming, and charge per delay are eliminated to reduce model complexity. A total of 219 data sets having five input features i.e., the number of holes, spacing, burden, specific drilling, and powder factor are used to develop the models. To predict rock fragmentation due to blasting in limestone quarry mines, both machine learning models (Random Forest Regression (Bagging), Support Vector Regression, and XG Boost Regression (Boosting)), as well as a deep learning model (Neural Network Regression), are applied to develop a model that can optimize the prediction of fragmentation. The Artificial neural network model optimization showed that the model with architecture 64-32-16-1 can perform well giving MSE (mean squared error) values of 41.32 and 28.59 on training and test data respectively. The R2 value for both training and test is 0.83. Random Forest regression is also performing well compared to SVR and XG boost with the MSE value 12.37 and 9.89 on training and testing data respectively. Here, the R2 value for both sets are 94%. Based on the permutation importance and Shapely plot values, the powder factor has the highest impact, and the burden has the lowest impact on fragmentation.

Variations in groundwater levels that are noticed in different parts of the world are typically the primary focus of research carried out in hydrology. The fluctuating levels of groundwater can be ascribed to several different factors, some of which include an increase in the demand for water, inappropriate irrigation practices, improper management of soil, and unregulated extraction from aquifers. It is necessary to effectively manage groundwater resources to have a trustworthy method of measuring and forecasting groundwater levels. The dynamics of groundwater are inherently uneven and difficult to understand. As a result, applying methods driven by data could potentially yield significant gains in hydrology. In this study, two data-driven models were utilized to estimate groundwater levels at a total of 4 monitoring wells located in Colorado state, USA. These models included ARIMA and ANN. The comparative analysis of the algorithms made use of a data set that had one month's worth of information from each of the years 1980 to 2019. The models were put through their paces, and the results of those tests were statistically and visually assessed. To evaluate the relative accuracy and precision of the models, the mean absolute error (MAE), mean square error (MSE), and root-mean-square error (RMSE) were used. When compared to ARIMA, the analysis demonstrates that ANN can produce the most accurate forecasts of groundwater levels in Colorado state USA.

Coastal wave modeling and forecasting are essential in oceanography, sustainable marine energy, and ocean engineering. Precise forecasting of wave speed and direction are crucial for offshore operations, marine energy, risk management, environmental management, coastal and sustainable maritime management. The coastline of Brunei Darussalam can generate between 15 and 126 Giga Watt of wave energy. In this experimental research, we used two numerical approaches, the finite difference, and spectral element methods, to model and simulate wave speed and direction. We calculated the mean error between numerical and analytical solutions. We proposed a novel, promising, univariate time series forecasting model by combining the Long Short-Term (LSTM) with KerasTuner hyperparameters tuning and optimization techniques. This method helps us to improve the accuracy and efficiency of time-series forecasting.. The experimental data was computed from high-precision Acoustic Doppler Current Profiler (ADCP) sensor data. This research is part of the preliminary feasibility analysis of wave energy production in Brunei Darussalam and net zero commitment for a sustainable environment. Seven independent forecast experiments were performed for wave speed and direction in degree and radian units for 1, 3, 6, 8, 10, 12, and 24 hours. Mean squared error (MSE) was adopted as a metric for both training and testing. The experimental results reveal that the wave speed forecast has the lowest MSE compared to direction, regardless of the unit of measure, but has a longer duration. In addition, the direction forecast in the degree unit has the lowest errors compared to the unit of radians; the latter has a longer running time than the former. The model has delivered optimal results throughout the experiments with minor training and test errors. We conducted a thorough evaluations on two benchmark time series datasets, which include the study dataset and air quality index dataset, to validate the performance of the proposed model with other models. The proposed model outperforms cutting-edge forecasting models, such as the conventional LSTM, ARIMA, and Prophet. The model has the slightest forecast error compared to the existing literature’s result.

Converting natural language into machine language that can be recognized by distributed systems is the core challenge of intelligent interactive interfaces and human-machine dialogue systems. The human-machine interface interaction of large distributed SCADA measurement and control system is tedious and the operation and maintenance cost is high, so it is significant to design an intelligent natural language interaction interface for distributed measurement and control system. In this paper, we design the intermediate language format of SCADA system, i.e., Key-value logic form, and then formulate the Text2SCADA framework and propose the TICS algorithm and SDPA algorithm, the former adopts the keyword extraction and cosine similarity optimization algorithm to complete the structured extraction of natural language for basic natural language instructions, and the latter adopts the keyword extraction and cosine similarity optimization algorithm to complete the structured extraction of natural language for relatively The latter one adopts the algorithm of dependent syntactic analysis for the structured representation of natural language instructions with relatively complex natural language instructions. Based on the above algorithms, a lightweight information extraction model based on DGCNN and probabilistic graph ideas is constructed, aiming to enhance the scientific generalization ability of the framework on unknown instruction sets. The experimental results show that the proposed intelligent natural language interface can better solve the human-machine interface interaction problem of distributed SCADA measurement and control system. The average accuracy, recall and F-value of instruction parsing reach 89.27\%, 89.28\% and 89.27\%, respectively. The average response time is 1.593 s. Especially, it provides a more convenient means of interaction for industrial and agricultural information control.

The continuously developing project aims to build an informatics system enabling analysis of spatial and temporal parameters of movement activities occurring in the sport of speed climbing. The monitoring system (climbing information speed system – CISS) is to be used for conducting comprehensive scientific research in the field of speed climbing. The system enables the evaluation of the training process of climbers at various levels of competition. The study analysis was based on video. The video recording with a camera positioned at a short distance (10 m) from the wall. The marker was positioned closest to the centre of mass (gravity) BMC. Results: development of a system for data collection and analysis of the climbing run based on video recording (application of the Kanade-Lucas-Tomasi (KLT) algorithm). Our results showed that used devices can measure a wide range of specific internal and external variables during speed climbing. Some of the analyzed parameters were significantly correlated with speed climbing time. These results could be a theoretical basis for future research and for training program’s preparation.

The agricultural sector in Palestine has a significant role in its economy. However, the production of this sector is affected by different kinds of plant diseases, specifically leaf diseases. Automatic agricultural leaf disease detection is essential for the early diagnosis and controls the overall health of fields. Image segmentation techniques, clustering, and deep learning are often used to detect diseased leaves. This study proposes a novel hybrid approach based on image classification. The hybrid approach combines the k-means clustering algorithm with Convolutional Neural Network (CNN), where k-means is used to detect the leaf’s infected area, then CNN is used for specifying the disease. We used the PlantVillage dataset for experimental verification as it contains several crops with different kinds of challenging diseases. We also examined the selection of optimal k-value using the Silhouette coefficient, Elbow method, and Kneedle Algorithm. The Silhouette technique was analyzed using three distance metrics; Euclidean, Manhattan, and Cosine. Its scores for the three-distance metrics were low, near-zero, and failed to produce the optimum k value. Besides, the Elbow method was complicated to use in image segmentation in terms of executing and visualizing the k value in its graph plot. Based on verification results, the Kneedle Algorithm produced better results in the consistency of choosing the optimal k value and showed superiority over other approaches. Therefore, the processed images were segmented with the k-means clustering algorithm with a Kneedle algorithm-based k value. Finally, a Convolutional Neural Network (CNN) is trained to classify the type of disease based on analyzing and testing leaf images. The hybrid model achieved high accuracy of 93.79% in disease identification, confirming the proposed model’s robustness.

Starting from a very simple economic scenario, we build on it a game and then we introduce a general strategy able to reduce a regression problem to an equivalent binary classification problem. This reduction scheme (that we call adaptive reduction or also dynamic reduction) can be also used to derive a new boosting algorithm for regression problems named bOOst^d. The bOOst^d algorithm is very simple to implement, and it can use any learning algorithm with no priori assumptions. We present a conjecture for bOOst^d performances, which ensures a little error on training set. More important we can also provide a very good theoretical upper bound for the generalization error. We give a set of preliminary experimental results that seems to confirm our conjecture for bOOst^d performances on training set and the theoretical assumptions for the generalization error. We also provide a possible justification of why boosting often does not overfit. Finally, we leave some open problems and argue that in the future an adaptive single boosting (with an unique code) algorithm for binary, multi class and regression problems can be derived.

In the leading/trailing edge’s adaptive machining of the near-net-shaped blade, a small portion of the theoretical part is retained for securing aerodynamic performance by manual work. However, this procedure is time-consuming and depends on the human experience. In this paper, we defined retained theoretical leading/trailing edge as the reconstruction area. To accelerate the reconstruction process, an anchor-free neural network model based on Transformer was proposed, named LETR (Leading/trailing Edge Transformer). LETR extracts image features from an aspect of mixed frequency and channel domain. We also integrated LETR with the newest meta-Acon activation function. We tested our model on the self-made dataset LDEG2021 on a single GPU and got an mAP of 91.9\%, which surpassed our baseline model, Deformable DETR by 1.1\%. Furthermore, we modified LETR’s convolution layer and named the new model after GLETR (Ghost Leading/trailing Edge Transformer) as a lightweight model for real-time detection. It is proved that GLETR has fewer weight parameters and converges faster than LETR with an acceptable decrease in mAP (0.1\%) by test results.

Deep learning (DL), a branch of machine learning (ML) and artificial intelligence (AI) is nowadays considered as a core technology of today's Fourth Industrial Revolution (4IR or Industry 4.0). Due to its learning capabilities from data, DL technology originated from artificial neural network (ANN), has become a hot topic in the context of computing, and is widely applied in various application areas like healthcare, visual recognition, cybersecurity, and many more. However, building an appropriate DL model is a challenging task, due to the dynamic nature and variations in real-world problems and data. Moreover, the lack of core understanding turns DL methods into black-box machines that hamper development at the standard level. This article presents a structured and comprehensive view on DL techniques including a taxonomy considering various types of real-world tasks like supervised or unsupervised. In our taxonomy, we take into account deep networks for supervised or discriminative learning, unsupervised or generative learning as well as hybrid learning and relevant others. We also summarize real-world application areas where deep learning techniques can be used. Finally, we point out ten potential aspects for future generation DL modeling with research directions. Overall, this article aims to draw a big picture on DL modeling that can be used as a reference guide for both academia and industry professionals.

Sleep deficiency impacts the quality of life and may have serious health consequences in the long run. Questionnaire-based subjective assessment of sleep deficiency has many limitations. On the other hand, objective assessment of sleep deficiency is challenging. In this study, we propose a polysomnography-based mathematical model for computing baseline sleep deficiency severity score and then investigated the estimation of sleep deficiency severity using features available only from wearable sensor data including heart rate variability and single-channel electroen-cephalography for a dataset of 500 subjects. We used Monte-Carlo Feature Selection (MCFS) and inter-dependency discovery for selecting the best features and removing multi-collinearity. For developing the Regression model we investigated both the frequentist and the Bayesian ap-proaches. An Artificial Neural Network achieved the best performance of RMSE = 5.47 and an R-squared value of 0.67 for sleep deficiency severity estimation. The developed method is com-parable to conventional methods of Functional Outcome of Sleep Questionnaire and Epworth Sleepiness Scale for assessing the impact of sleep apnea on sleep deficiency. Moreover, the results pave the way for reliable and interpretable sleep deficiency severity estimation using a wearable device in Smart Health.

Deep learning (DL), which is originated from an artificial neural network (ANN), is one of the major technologies of today's smart cybersecurity systems or policies to function in an intelligent manner. Popular deep learning techniques, such as Multi-layer Perceptron (MLP), Convolutional Neural Network (CNN or ConvNet), Recurrent Neural Network (RNN) or Long Short-Term Memory (LSTM), Self-organizing Map (SOM), Auto-Encoder (AE), Restricted Boltzmann Machine (RBM), Deep Belief Networks (DBN), Generative Adversarial Network (GAN), Deep Transfer Learning (DTL or Deep TL), Deep Reinforcement Learning (DRL or Deep RL), or their ensembles and hybrid approaches can be used to intelligently tackle the diverse cybersecurity issues. In this paper, we aim to present a comprehensive overview from the perspective of these neural networks and deep learning techniques according to today's diverse needs. We also discuss the applicability of these techniques in various cybersecurity tasks such as intrusion detection, identification of malware or botnets, phishing, predicting cyber-attacks, e.g. denial of service (DoS), fraud detection or cyber-anomalies, etc. Finally, we highlight several research issues and future directions within the scope of our study in the field. Overall, the ultimate goal of this paper is to serve as a reference point and guidelines for the academia and professionals in the cyber industries, especially from the deep learning point of view.

Luminescence-based sensors for measuring oxygen concentration are widely used both in industry and research due to the practical advantages and sensitivity of this type of sensing. The measuring principle is the luminescence quenching by oxygen molecules, which results in a change of the luminescence decay time and intensity. In the standard approach, this change is related to an oxygen concentration using the Stern–Volmer equation. This equation, which in most of the cases is non-linear, is parametrized through device-specific constants. Therefore, to determine these parameters every sensor needs to be precisely calibrated at one or more known concentrations. This work explores an entirely new artificial intelligence approach and demonstrates the feasibility of oxygen sensing through machine learning. The specifically developed neural network learns very efficiently to relate the input quantities to the oxygen concentration. The results show a mean deviation of the predicted from the measured concentration of 0.5 % air, comparable to many commercial and low-cost sensors. Since the network was trained using synthetically generated data, the accuracy of the model predictions is limited by the ability of the generated data to describe the measured data, opening up future possibilities for significant improvement by using a large number of experimental measurements for training. The approach described in this work demonstrates the applicability of artificial intelligence to sensing technology and paves the road for the next generation of sensors.

Non-invasive, electroencephalography (EEG)-based brain-computer interfaces (BCIs) on motor imagery movements translate the subject’s motor intention into control signals through classifying the EEG patterns caused by different imagination tasks, e.g. hand movements. This type of BCI has been widely studied and used as an alternative mode of communication and environmental control for disabled patients, such as those suffering from a brainstem stroke or a spinal cord injury (SCI). Notwithstanding the success of traditional machine learning methods in classifying EEG signals, these methods still rely on hand-crafted features. The extraction of such features is a difficult task due to the high non-stationarity of EEG signals, which is a major cause for the stagnating progress in classification performance. Remarkable advances in deep learning methods allow end-to-end learning without any feature engineering, which could benefit BCI motor imagery applications. We developed three deep learning models: 1) a long short-term memory (LSTM); 2) a proposed spectrogram-based convolutional neural network model (pCNN); and 3) a recurrent convolutional neural network (RCNN), for decoding motor imagery movements directly from raw EEG signals without (manual) feature engineering. Results were evaluated on our own, publicly available, EEG data collected from 20 subjects and on an existing dataset known as 2b EEG dataset from "BCI Competition IV". Overall, better classification performance was achieved with deep learning models compared to state-of-the art machine learning techniques, which could chart a route ahead for developing new robust techniques for EEG signal decoding. We underpin this point by demonstrating the successful real-time control of a robotic arm using our CNN based BCI.

Vibration measurement serves as the basis for various engineering practices such as natural frequency or resonant frequency estimation. As image acquisition devices become cheaper and faster, vibration measurement and frequency estimation through image sequence analysis continue to receive increasing attention. In the conventional photogrammetry and optical methods of frequency measurement, vibration signals are first extracted before implementing the vibration frequency analysis algorithm. In this work, we demonstrated that frequency prediction can be achieved using a single feed-forward convolutional neural network. The proposed method is verified using a vibration signal generator and excitation system, and the result obtained was compared with that of an industrial contact vibrometer in a real application. Our experimental results demonstrate that the proposed method can achieve acceptable prediction accuracy even in unfavorable field conditions.

Smart buildings concept aims at the use of the smart technology to reduce energy consumption as well as improvement of the comfort conditions and users’ satisfaction. It is based on the use of smart sensors to follow both outdoor and indoor conditions as well as software for the control of comfort and security devices. The optimal control of the energy devices requires software for indoor temperature forecasting. This paper presents an ANN – based model for the indoor temperature forecasting. The model is developed using data recoded in an old building of the engineering school Polytech’Lille. Data covered both indoor and outdoor conditions. Analysis of the relevance of the input parameters allowed to develop a simplified forecasting model of the indoor temperature that uses only the outdoor temperature as well as the history of the façade temperature as input parameters. The paper presents successively, data collection, the ANN concept used in the temperature forecasting, and finally the ANN model developed for the façade and indoor forecasting. It shows that an ANN-based model using outdoor and façade temperature sensors provides a good forecasting of the indoor temperature. This model could be used for the optimal control of buildings energy devices.

The COVID-19 was described as a respiratory illness, however further studies recognize it as a complex heterogeneous multisystemic disorder. Global efforts have been proposed to combat COVID-19, emerging diverse therapeutic options, in which discovering new drug therapies, development of vaccines and drug repurposing have been considered the most promising approaches to fight the virus. This study aimed to repurpose known drugs for use against the COVID-19, finding better therapeutic options. Seventeen biological databases were used in this study. The genetic algorithm (GA) was performed for a set of drug target classes and COVID-19 proteins as input, whose drug candidates are obtained according to the target similarities found in the target-target similarity predictive network, resulting in a drug-target interaction network. Thus, recommended drugs correspond to the union of the drug subsets found during each GA execution. Twenty-eight drugs were indicated to be the best therapeutic targets for the virus, in special, the Cyclosporine drug was administered as adjuvant to steroid treatment for COVID-19 patients which showed positive outcomes, reducing mortality in moderate and severe cases. The drugs found have used to treat other diseases, evidencing that the COVID-19 is a multisystemic disorder and suggests that the viruses’ mechanism of action presents some comorbidity with other human diseases. Evidence shows that the drugs found in this research might act together to fight the virus in a broader fashion, however further studies including in vitro and in vivo experiments are needed to find the best combination of these drugs.

The RPL protocol is a routing protocol for low power and lossy networks. In such a network, energy is a very scarce resource; so many studies are focused on minimizing global energy consumption. End-to-end latency is another important performance indicator of the network, but existing research tends to focus more on energy consumption and ignore the end-to-end delay of data transmission. In this paper, we propose a kind of energy equalization routing protocol to maximize the surviving time of the restricted nodes so that the energy consumed by each node is close to each other. At the same time, a multi-path forwarding route is proposed based on the cache utilization. The data is sent to the sink node through different parent nodes at a certain probability, not only by selecting the preferred parent node, thus avoiding buffer overflow and reducing end-to-end delay. Finally, the two algorithms are combined to accommodate different application scenarios. The experimental results show that the proposed three improved schemes improve the reliability of the routing, extend the lifetime of the network, reduce the end-to-end delay, and reduce the number of DAG reconfiguration.

With the continuous development of network technology, complex network systems generate massive unbalanced attack traffic. Due to the severe imbalance in the quantities of normal samples and attack samples, as well as among different types of attack samples, intrusion detection systems suffer from low detection rates for rare class attack data. In this paper, we propose a geometric synthetic minority oversampling technique based on optimized kernel density estimation algorithm. This method can generate diverse rare class attack data by learning the distribution of rare class attack data while maintaining similarity with the original sample features. Meanwhile, the balanced data is input to a feature extraction module built upon multiple denoising autoencoders, reducing information redundancy in high-dimensional data and improving the detection performance for unknown attacks. Subsequently, a soft voting ensemble learning technique is utilized for multi-class anomaly detection on the balanced and dimensionally reduced data. Finally, an intrusion detection system is constructed based on data preprocessing, imbalance handling, feature extraction, and anomaly detection modules, and validated on the NSL-KDD and N-BaIoT datasets. Comparative experiments with baseline models and other state-of-the-art methods demonstrate that the proposed system improves the detection rate of rare class attack data. Furthermore, it achieves a good overall detection rate on the Internet of Things dataset (N-BaIoT), indicating its strong applicability.

In recent years web applications and on-line business transactions have grown many folds. Consequently, also cyberattacks have increased and represent a serious threat to the pervasive digital services upon which our society relies. To mitigate cyberattacks, many countermeasures are deployed on computing nodes (e.g., anti-malware software) as well as on network devices to detect and possibly block malicious packets in transit; these monitoring devices broadly go under the name of firewalls. Firewalls are designed according to two main architectural approaches: software running on a standard or embedded computer, or purposedly designed hardware, e.g., ASICs. Software-based solutions have the advantage of high flexibility and can be ported on easily upgradable hardware. However, hardware implementation represents the only viable solution for high data rates. On the market, very fast devices of the latter kind are available, but their cost is typically very high, especially considering that their ultra-optimized design makes updating them very difficult, with the consequence of a rather short lifespan. As a more balanced alternative, we wanted to investigate the use of an FPGA architecture, , which is significantly easier to update than custom-built chips, and features low-latency and high-throughput characteristics concurrently, making it preferrable to other programmable systems based on GPUs or microcontrollers. In this paper a packet sniffer that has been designed on FPGA with a 1 Gbit/s data transfer rate is presented. The system is implemented on the FPGA development board KC705 by Xilinx, can analyze Ethernet frames, checking the frame fields against a set of rules defined by the user and calculates statistics of the received Ethernet frames over time. The designed packet sniffer has been successfully tested both with Ethernet frames ad hoc generated using a packets generator, and with real web traffic by connecting the packet sniffer to the internet.

Geodetic networks are essential for most geodetic, geodynamics and civil projects, such as monitoring the position and deformation of man-made structures, monitoring the crustal deformation of the Earth, establishing and maintaining a geospatial reference frame, mapping, civil engineering projects and so on. Before the installation of geodetic marks and gathering of survey data, geodetic networks need to be designed according to the pre-established quality criteria. In this study, we present a method for designing geodetic networks based on the concept of reliability. We highlight that the method discards the use of the observation vector of Gauss-Markov model. In fact, the only needs are the geometrical network configuration and the uncertainties of the observations. The aim of the proposed method is to find the optimum configuration of the geodetic control points so that the maximum influence of an outlier on the coordinates of the network is minimum. Here, the concept of Minimal Detectable Bias defines the size of the outlier and its propagation on the parameters is used to describe the external reliability. The proposed method is demonstrated by practical application of one simulated levelling network. We highlight that the method can be applied not only for geodetic network problems, but also in any branch of modern science.

To drive safely, the driver must be aware of the surroundings, pay attention to the road traffic, and be ready to adapt to new circumstances. Most studies on driving safety focus on detecting anomalies in driver behavior and monitoring the cognitive capabilities of drivers. In our study, we proposed a classifier for basic activities in driving a car, based on a similar approach that could be applied to the recognition of basic activities in daily life, that is, using electrooculographic (EOG) signals and a one-dimensional convolutional neural network (1D CNN). Our classifier achieved an accuracy of 80% for the 16 primary and secondary activities. The accuracy related to primary activities in driving, including crossroad, parking, roundabout was 97.9%, 96.8%, 97.4%, and 99.5%, respectively. The F1 score for secondary driving actions (0.99) was higher than for primary driving activities (0.93–0.94). Furthermore, using the same algorithm, it was possible to distinguish four secondary activities related to activities of daily life and secondary when driving a car.

Autonomous mobile robots are essential to the industry, and human-robot interactions are becoming more common nowadays. These interactions require that the robots navigate scenarios with static and dynamic obstacles in a safely manner, avoiding collisions. This paper presents a physical implementation of a method for dynamic obstacle avoidance using a Long Short-Term Memory (LSTM) neural network that obtains information from the mobile robot’s LiDAR for it to be capable of navigating through scenarios with static and dynamic obstacles while avoiding collisions and reaching its goal. The model is implemented using a TurtleBot3 mobile robot within an OptiTrack Motion Capture (MoCap) system for obtaining its position at any given time. The user operates the robot through these scenarios, recording its LiDAR readings, target point, position inside the MoCap system, and its linear and angular velocities, all of which serve as the input for the LSTM network. The model is trained on data from multiple user-operated trajectories across five different scenarios, outputting the linear and angular velocities for the mobile robot. Physical experiments prove that the model is successful in allowing the mobile robot to reach the target point in each scenario while avoiding the dynamic obstacle, with a validation accuracy of 98.02%.

The forecast of electricity demand has been a recurrent research topic for decades, due to its economical and strategic relevance. Several Machine Learning (ML) techniques have evolved in parallel with the complexity of the electric grid. This paper reviews a wide selection of approaches that have used Artificial Neural Networks (ANN) to forecast electricity demand, aiming to help newcomers and experienced researchers to appraise the common practices and to detect areas where there is room for improvement in the face of the current widespread deployment of smart meters and sensors, which yields an unprecedented amount of data to work with. The review looks at the specific problems tackled by each one of the selected papers, at the results attained by their algorithms, and at the strategies followed to validate and compare the results. This way, it is possible to highlight some peculiarities and algorithm configurations that seem to consistently outperform others in specific settings.

The sustainability of the planet and its inhabitants is in dire danger and is among the highest priorities on global agendas such as the Sustainable Development Goals (SDGs) of the United Nations (UN). Solar energy -- among other clean, renewable, and sustainable energies -- is seen as essential for environmental, social, and economic sustainability. Predicting solar energy accurately is critical to increasing reliability and stability, and reducing the risks and costs of the energy systems and markets. Researchers have come a long way in developing cutting-edge solar energy forecasting methods. However, these methods are far from optimal in terms of their accuracies, generalizability, benchmarking, and other requirements. Particularly, no single method performs well across all climates and weathers due to the large variations in meteorological data. This paper proposes SENERGY (an acronym for Sustainable Energy), a novel deep learning-based auto-selective approach and tool that, instead of generalising a specific model for all climates, predicts the best performing deep learning model for GHI forecasting in terms of forecasting error. The approach is based on carefully devised deep learning methods and feature sets through an extensive analysis of deep learning forecasting and classification methods using ten meteorological datasets from three continents. We analyse the tool in great detail through a range of metrics and methods for performance analysis, visualization, and comparison of solar forecasting methods. SENERGY outperforms existing methods in all performance metrics including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Forecast Skills (FS), Relative Forecasting Error, and the normalised versions of these metrics. The proposed auto-selective approach can be extended to other research problems such as wind energy forecasting and predict forecasting models based on different criteria (in addition to the minimum forecasting error used in this paper) such as the energy required or speed of model execution, different input features, different optimisations of the same models, or other user preferences.

Shape-shifting polymers are widely used in various fields such as intelligent switches, soft robots and sensors, which require both multiple stimulus response functions and qualified mechanical strength. In this study, a novel near infrared light (NIR) responsible shape-shifting hydrogel system was designed and fabricated, through embedding vinylsilane modified carbon nanotubes (CNTs) into particle-double network (P-DN) hydrogels by micellar copolymerization. The dispersed brittle Poly(sodium 2-acrylamido-2-methylpropane-1-sulfonate) (PNaAMPS) network of the microgels can serves as sacrificial bonds to toughen the hydrogels, and the CNTs endow it NIR photothermal conversion ability. The results show that the CNTs embedded P-DN hydrogels present excellent mechanical strength, i. e. fracture strength of 312 kPa and fracture strain of 357%. Moreover, an asymmetric bilayer hydrogel, where the active layer contains CNTs, can achieve 0°–110° bending deformation within 10 min under NIR irradiation and can realise complex deformation movement. This study provides a theoretical and experimental basis for the design and manufacture of photoresponsive soft actuators.

Although there is a vast body of neuroscientific research on cognitive behavioral therapy (CBT), many therapists are confused about how to incorporate the evidence into their clinical practice. This narrative review translates neuroscientific research on CBT for depression into a therapist-friendly formulation model. Four brain networks that are particularly relevant to depression, including the Central Executive Network, Default mode Network, Salience Network, and Affective Network, were identified based on their interactions and described in relation to depressive symptoms. Additionally, it clarified how each CBT technique works on a brain map. Assessment, intervention selection, and patient psychoeducation based on this model will facilitate the clinical application of neuroscience in CBT.

The exponential growth of digital media content has introduced new challenges in managing and classifying internet traffic. Digital media traffic is composed of various applications such as video, audio, social media, and search, and its data structure is complex, incorporating a vast array of features. The classification of traffic data is a crucial aspect of internet traffic management and network security, and it forms the basis for several scenarios, including content distribution, advertising recommendations, and data analysis. Traditional classification methods rely mainly on deep packet inspection and port-based techniques, which have become increasingly ineffective due to the rapid evolution of network traffic. To address this issue, this study proposes a machine learning-based traffic classification method aimed at enhancing the accuracy and efficiency of digital media traffic classification to meet the current needs of traffic management and network security. The paper also analyzes and evaluates the classification effect and prediction capability of various algorithms under different training set sizes to validate the feasibility and effectiveness of the proposed method. The result demonstrates that the neural network algorithm has superior classification and prediction capabilities compared to the decision tree and support vector machine algorithms. Furthermore, our proposed method achieves the highest accuracy of 96.88% with a large training sample of 40,000 data streams, proving its superiority in handling high-dimensional data and complex datasets. The research results are significant for the development of digital media traffic classification and prediction methods and are expected to be applied in practical scenarios.

In the past few decades, the transmission of data over an unsecure channel resulted in an increased rate of hacking. Therefore, the requirement to make multimedia data more secure is increasing day by day. Numerous algorithms are developed to improve efficiency and robustness. In this article, a novel and secure image encryption algorithm is presented. It is based on a modified chaotic logistic map (CLM) that provides the advantage of having less computational time to encrypt an input image. The encryption algorithm is based on Shannon’s idea of using substitution-permutation and one-time-pad network to achieve ideal secrecy. The CLM is used for substitution and permutation to improve randomness and increase dependency on the encryption key. Various statistical tests are conducted, such as Key Space Analysis, Complexity Analysis, Sensitivity Analysis, Strict Avalanche Criteria (SAC), Histogram Analysis, Entropy Analysis, Mean of Absolute Deviation (MAD) Analysis, Correlation Analysis, Contrast Analysis, and Homogeneity to give a comparative analysis of the proposed algorithm and verify its security. As a result of various statistical tests, it is evident that the proposed algorithm is more efficient and robust as compared to previous ones.

This paper presents a novel approach to sentiment analysis specifically customized for predicting stock market movements, bypassing the need for external dictionaries which are often unavailable for many languages. Our methodology directly analyzes textual data, with a particular focus on context-specific sentiment words within neural network models. This specificity ensures our sentiment analysis is both relevant and accurate in identifying trends in the stock market. We employ sophisticated mathematical modeling techniques to enhance both the precision and interpretability of our models. Through meticulous data handling and advanced machine learning methods, we leverage large datasets from Twitter and financial markets to examine the impact of social media sentiment on financial trends. We achieved an accuracy exceeding 75%, highlighting the effectiveness of our modeling approach, which we further refined into a convolutional neural network model. This achievement contributes valuable insights into sentiment analysis within the financial domain, thereby improving the overall clarity of forecasting in this field.

Convolutional neural networks (CNN) have attracted much attention as a commonly used method for hyperspectral image (HSI) classification in recent years, however, CNNs can only be applied to Euclidean data and have limitations in dealing with relationships due to the limitations of local feature extraction. However, each pixel of a hyperspectral image contains a set of spectral bands that are correlated and interact with each other, and the methods used to process Euclidean data cannot effectively obtain these correlations. In contrast, the graph convolutional network (GCN) can be used in non-Euclidean data, but usually leads to oversmoothing and ignoring local detail features due to the need for superpixel segmentation processing to reduce computational effort. To overcome the above problems, we constructed a network a fusion network based on GCN and CNN, which contains two branches: a graph convolutional network based on superpixel segmentation and a convolutional network with added attention mechanism. The graph convolu-tional branch can extract the structural features and capture the relationships between the nodes, and the convolutional branch can extract the detailed features in the local fine region. Owing to the fact that the features extracted from the two branches are different, the classification performance can be improved by fusing the complementary features extracted from the two branches. To vali-date the proposed algorithm, experiments were conducted on three widely used datasets, namely Indian Pines, Pavia University, and Salinas, and the overall accuracy of 98.78% was obtained in the Indian Pines dataset, and the overall accuracy of 98.99% and 98.69% was obtained in the other two datasets. The results showed that the proposed fusion network can obtain richer features and achieve high classification accuracy.

Unmanned aerial vehicle UAV has become an important tool for emergency communication. When disaster happens, the roads in the disaster area always are seriously damaged, and theground infrastructure communication facilities are damaged. Under emergency conditions, UAV can go into key areas as communication nodes to provide communication services for users in the area.But it's difficult for UAV to cover the entire disaster area and the efficient rescue time will be delayed. Therefore, how to make the UAV accurately perceive the regional situation and deployed in the right position is worth studying. In this paper, a virtual simulation environment is established reasonably and deep reinforcement learning algorithm is used to train UAV agents. At the same time, a series of problems such as sparse reward and long training time are still not solved in the development of reinforcement learning algorithms,and the reward function is designed to improve training efficiency. First of all, we set a specific mountain emergency communication scenario, and combined with the specific application of the UAV to carry out virtual simulation, and build a virtual environment. Furthermore, an additional shaped reward function is designed to address the sparse reward problem. Through the improvement of deep Q-learning network DQN algorithm and the reward design based on potential function, the final evaluation index is improved, and the effectiveness of the algorithm is verified. The experimental results demonstrate our work can effectively shorten the training time and increase convergencerate.

This study aims to develop novel automated computer vision algorithms and systems for component replacement inspection for Printed Circuit Boards (PCBs). The proposed algorithms are able to identify locations as well as sizes of different components. They are object detection algorithms based on keypoints of the target components. The algorithms can be implemented as neural networks consisting of two portions: frontend networks and backend networks. The frontend networks are used for the feature extractions of input images. The backend networks are adopted for producing component inspection results. Each component class can has its own frontend and backend networks. In this way, the neural model for the component class can be effectively reused for different PCBs. To reduce the computation time for the inference of the networks, different component classes can share the same frontend networks. A two-stage training process is proposed for effectively exploring features of different components for accurate component inspection. The proposed algorithm has the advantages of the simplicity for training data collection, high accuracy for defect detection, and high reusability and flexibility for online inspection. The algorithm is an effective alternative for the automated inspection in smart factory with growing demand for product quality and diversification.

For the Remotely Piloted Aircraft Systems (RPAS) market to continue its current growth rate, cost-effective "Detect and Avoid" systems that enable safe beyond visual line of sight (BVLOS) operations are critical. We propose an audio-based "Detect and Avoid" system, composed of microphones and an embedded computer, which performs real-time inferences using a sound event detection (SED) deep learning model. Two state-of-the-art SED models, YAMNet and VGGish, are fine-tuned using our dataset of aircraft sounds and their performances are compared for a wide range of configurations. YAMNet, whose MobileNet architecture is designed for embedded applications, outperformed VGGish both in terms of aircraft detection and computational performance. YAMNet's optimal configuration, with > 70% true positive rate and precision, results from combining data augmentation and undersampling with the highest available inference frequency (i.e. 10 Hz). While our proposed "Detect and Avoid" system already allows the detection of small aircraft from sound in real time, additional testing using multiple aircraft types is required. Finally, a larger training dataset, sensor fusion, or remote computations on cloud-based services could further improve system performance.

When extract building from high resolution remote sensing image with meter/sub-meter accuracy, the shade of trees and interference of roads are the main factors of reducing the extraction accuracy. Proposed a Bayesian Convolutional Neural Networks(BCNET) model base on standard fully convolutional networks(FCN) to solve these problems. First take building with no shade or artificial removal of shade as Sample-A, woodland as Sample-B, road as Sample-C. Set up 3 sample libraries. Learn these sample libraries respectively, get their own set of feature vector; Mixture Gauss model these feature vector set, evaluate the conditional probability density function of mixture of noise object and roofs; Improve the standard FCN from the 2 aspect:(1) Introduce atrous convolution. (2) Take conditional probability density function as the activation function of the last convolution. Carry out experiment using unmanned aerial vehicle(UVA) image, the results show that BCNET model can effectively eliminate the influence of trees and roads, the building extraction accuracy can reach 97%.

Continuous object tracking in WSNs, such as monitoring of mud-rock flows, forest fires etc., is a challenging task due to characteristic nature of continuous objects. They can appear randomly in the network field, move continuously, and can change in size and shape. Monitoring such objects in real-time generally require tremendous amount of messaging between sensor nodes to synergistically estimate object’s movement and track its location. In this paper, we propose a novel twofold-sink mechanism, comprising of a mobile and a static sink node. Both sink nodes gather information about boundary sensor nodes, which is then used to uniformly distribute energy consumption across all network nodes, thus helping in saving residual energy of network nodes. Numerous object tracking schemes, using mobile sink, have been proposed in the literature. However, existing schemes employing mobile sink cannot be applied to track continuous objects, because of momentous variation of network topology. Therefore, we present in this paper a mechanism, transformed from K-means algorithm, to find the best sensing location of the mobile sink node. It helps to reduce transmission load on the intermediate network nodes situated between static sink node and the ordinary network sensing nodes. The simulation results show that the proposed scheme can distinctly improve life time of the network, compared to one-sink protocol employed in continuous object tracking.

At the end of 2018, a high school student asked a question in Zhihu community, claiming that he had proved Goldbach's conjecture. The problem caused an explosive reaction and a large number of users participated in the discussion. And has caused the widespread influence. On January 1, 2019, the questioner issued his "proof". His proof was soon proved wrong. The heated discussion caused by the incident contains a lot of information of social science analysis value. Therefore, we follow up the event in the first time and build a time series dataset for the event. Taking the questioner's "proof" as the dividing line, all the answers, comments, sub comments and user information of writing these texts before and after two days were recorded. This series of temporal information can reflect the dynamic features of the interaction between user opinions, and the impact of exogenous shocks (proof release) on community opinions. The dataset can be used not only for the demonstration of various social network analysis algorithms, but also for a series of natural language processing tasks such as fine-grained sentiment analysis for long texts, as well as multimodal tasks combining natural language processing and social network analysis. This paper introduces the characteristics and structure of the dataset, shows the visualization effect of social network, and uses the dataset train the benchmark model of emotion analysis.

Epilepsy is a neurological disorder that affects approximately 1% of the world's population. To diagnose and estimate the occurrence of epilepsy, the analysis of recorded brain activity is performed by a neurologist, which is not only time-consuming and tedious but also occasionally accompanied by human error. Therefore, in recent decades, researchers have aimed to unravel an approach for designing and building an automated method for diagnosing and estimating the occurrence of epilepsy. Accordingly, the present study proposed two new-fangled ways based on brain signals and a convolutional neural network (CNN). Moreover, this research implements a CNN with a sequential three-layer structure. Numerous experiments were performed, and the accuracy of estimating epilepsy using the developed methods was achieved at 95% without feedback and 97% with feedback. The proposed methods were proven to be more accurate than the previous techniques and can be employed as a physician's assistant once entering the field of operation.

Background: Citicola tubulosa (CT) is a traditional Chinese medicine that has been found to have hepatoprotective effect，but the underlying mechanisms against acute alcoholic liver injury (ALI) remain unclear. Objective：This work aims to predict the molecular mechanisms of CT in treating ALI using network pharmacology and molecular docking. Methods: Targets of CT and ALI were screened and collected from TCMSP, Genecard, OMIM and other databases. The effective ingredient target network was constructed using Cytoscape platform. The core targets were obtained by constructing protein interaction network. The biological processes and pathways of CT were analyzed using the DAVID platform. The effective components of CT were docked with core targets using AutoDock software, and visualized by PyMOL software. Results: There were seven main active components, 89 drug targets, 6291 disease targets and 32 pathways. The top three docking results were hydrogen bonding between stigmasterol and EGFR, β⁃ sitosterol and ALB, stigmasterol and BCL2. Conclusion: The effective components of CT (e.g., quercetin, stigmasterol, crocetin) may regulate core targets via multiple pathways to alleviate ALI. The use of network pharmacology and molecular docking technology can aid in revealing effective components and underlying molecular mechanisms of CT.

In this paper, we propose a robust video understanding model for activity recognition by learning the actor’s pair-wise correlations and relational reasoning, exploiting spatial and temporal information. In order to measure the similarity between the pair appearances and construct an actor relations map, the Zero Mean Normalized Cross-Correlation (ZNCC) and the Zero Mean Sum of Absolute Differences(ZSAD) is proposed to allow the Graph Convolution Network (GCN) to learn how to distinguish group actions. We recommend that MNASNet be used as the backbone to retrieve features. Experiments show a 38.50% and 23.7% reduction in training time in the 2-stage training process along with a 1.52% improvement in accuracy against traditional methods.

With prosperous development and successful application of deep learning technologies in the field of remote sensing, numerous deep-learning based methods have emerged for synthetic aperture radar (SAR) automatic target recognition (ATR) tasks over the past few years. Generally, most deep-learning based methods can achieve outstanding recognition performance in condition that an abundant of labeled samples are available to train the model. However, in real application scenarios, it is difficult and costly to acquire and to annotate abundant SAR images due to the imaging mechanism of SAR, which poses a big challenge to existing SAR ATR methods. Thereby, SAR target recognition under the situation of few-shot, where only a scarcely few of labeled samples are available, is a fundamental problem that needs to be solved. In this paper, we put forward a new method named enhanced prototypical network with customized region-aware convolution (CRCEPN) to specially tackle the few-shot SAR ATR tasks. To be specific, We first develop a feature extraction network based on a customized and region-aware convolution, which can adaptively adjust convolutional kernels and their receptive fields according to each SAR image’s own characteristics as well as the semantical similarity among spatial regions, thus augmenting its capability to extract more informative and discriminative features. To achieve accurate and robust target identity prediction under the few-shot condition, we then propose an enhanced prototypical network which can improve the representation ability of the class prototype by properly making use of training and test samples together, thus effectively raising the classification accuracy. Meanwhile, a new hybrid loss is designed to learn a feature space with both inter-class separability and intra-class tightness as much as possible, which can further upgrade the recognition performance of the proposed method. Experiments performed on the moving and stationary target acquisition and recognition (MSTAR) dataset and the OpenSARShip dataset demonstrate that the proposed method is competitive with some state-of-the-art methods for few-shot SAR ATR tasks.

Machine-learning methods used for classification are challenged by the class imbalance problem, where a certain class is underrepresented. Over/under-sampling the majority/minority class observations or model selection for ensemble methods alone might not be effective if the class imbalance ratio is very high. To address this concern, a novel method is presented for generating synthetic data for minority class observations in conjunction with an optimal tree ensemble classifier (OTEC). This novel synthetic method first generates minority class instances to balance the training data. Then it applies OTEC, where models are selected based on their performance using out-of-bag observations and training subsamples. A total of 20 benchmark problems on binary classification with moderate to extreme class imbalance are used to assess the efficacy of the proposed method against other well-known methods, including optimal tree ensemble, smote random forest, under-sampling random forest, oversampling random forest, k-nearest-neighbor, support vector machine, tree, and artificial neural network, using classification error rate, sensitivity, specificity, precision, recall and F1-score as performance metrics to assess the efficacy of the proposed method. The analyses presented in this study revealed that the proposed method based on data balancing and model selection yielded better results than other methods.

Aiming at solving difficulties related to aero-engine classification and identification, the infrared spectra of aero-engine hot jets are measured by telemetry Fourier transform infrared spectrometer, and a six types aero-engine hot jet spectrum data set is generated. The spectrum data set were randomly sampled at a ratio of 8:1:1 to generate training set, validation set and prediction set. In this paper, a Peak Finding Siamese Convolutional Neural Network (PF-SCNN) is designed to match and classify spectrum data. In the first place, the matching pairs of training set and validation set are made independently with the labels of positive and negative samples creating. The spectrum feature extraction and distance similarity calculation are carried out by the Siamese convolutional neural network (SCNN). When entering the prediction phase, the prediction set and training set are integrated to predict in the trained model. The prediction set labels are ultimately determined by matching them with the training set labels having the highest similarities to ensure accurate predictions. To improve the operation efficiency of the SCNN, a peak finding method is introduced to extract the spectrum peaks. Peak positions of high frequency are counted, while the peak data are extracted. The peak data are used to take the place of the original data for training and prediction. The performance measures are defined as Accuracy, Precision, Recall, Confusion matrix and F1-score, and the prediction accuracy was up to 99%. The experimental results indicate that the proposed PF-SCNN is suitable for our spectral dataset and can complete the task of infrared spectrum classification of aero-engine hot jets.

This study explores the prediction of concrete compressive strength using machine learning models, aiming to overcome the time-consuming and complex nature of conventional methods. Four models—Artificial Neural Network, Multiple Linear Regression, Support Vector Machine, and Regression Tree—are employed and compared for performance. Evaluation metrics such as Mean Absolute Deviation, Root Mean Square Error, Coefficient of Correlation, and Mean Absolute Percentage Error are used. After preprocessing 1030 samples, the dataset is split into 70% for training and 30% for testing. The ANN model, further divided into training, validation (15%), and testing (15%), outperforms others in accuracy and efficiency. This outcome streamlines compressive strength determination in the construction industry, saving time and simplifying the process.

There are several challenges on how to attain energy efficiency while maintaining balance among factors affecting energy consumption such as power rating and temperature setpoints of cooling units, room temperature and humidity, and indoor air quality (IAQ). A real-time energy and IAQ monitoring system were installed in an educational building to profile the operational power consumption of inverter-based air condition unit (ACU) installed in each room. Polynomial curve and neural fitting regression analysis were applied to the real-time power consumption, indoor temperature and humidity, and carbon dioxide (CO2) level data. The derived models were able to provide the stabilizing indoor thermal and air conditions of the room to reach steady state ACU power consumption. These collected data and calculated parameters can be used to define rules in automating control of cooling appliances for an efficient energy utilization. The regression approaches, using real-time data, have determined the influence of indoor heat conditions and carbon dioxide levels on ACU power consumption. These parameters were utilized to establish consistent values for temperature, humidity, and carbon dioxide levels under stable settings of inverter-based air conditioning units.

This work studies the discontinuities features of sedimentary flysch materials in a 100 km2 area belonging to the Basque Arc. Such materials are common in this Spanish Alpine region located in the north of the Iberian Peninsula. A total of 33 outcrops are investigated by an intensive geotechnical investigation including geomechanical stations, boreholes and mechanical laboratory tests. Two flysch units are characterized: the Upper Aguinaga Formation or siliciclastic flysch, and the Lower Itziar Formation or calcareous flysch. Differences between both flysch formation are found. Joints in the siliciclastic flysch formation present an undulated roughness, with a spacing narrower and a persistence lower than in the calcareous flysch formation, which exhibits higher friction angles, although roughness is essentially planar. In addition, the potential of using Artificial Intelligence (AI) techniques, particularly Artificial Neural Networks and Support Vector Machine, to estimate the Geological Strength Index (GSI) from the Rock Quality Design (RQD) and some discontinuities features (spacing, persistence, aperture and roughness) is investigated. AI techniques are found to be satisfactory, being the Support Vector Machine with a linear kernel the technique which achieves the best performance.

The models of partial least squares (PLS) regression and artificial neural network (ANN) for evaluation of texture properties of cooked germinated brown rice (GBR) using the Fourier transform near infrared (NIR) spectra of uncooked whole grain combined with data separation methods, spectral pretreatment methods were investigated in this study. The ANN was outperformed in evaluation of hardness by back extrusion test of cooked GBR using the smoothing combined with standard normal variate pretreated NIR spectra in the range of 12,500-4,000 cm−1 of 188 whole grain samples where the calibration sample set was separated from prediction set by Kennard-Stone method where the best ANN model for hardness from hidden layers of 25 and 8 iteration time provided R2, r2, RMSEC, RMSEP, Bias and RPD of 0.9987, 0.9447, 0.1021 N, 0.7699 N, 0.0216 N and 4.3 respectively. The PLS regression of 64 sample KDML GBR group and 64 sample various variety GBR group, provided models for the hardness of the former and the toughness of the latter which developed by using 7506−5446.3, 4605.4−4242.9 cm−1 which included the amylose vibration band of 6834 cm−1 while toughness model was from 9403.8-6094.3 cm−1 where included 6834 and 8316 cm−1 vibration bands of amylose which influenced the texture of cooked rice. The PLS regression models for hardness and toughness were with the r2 of 0.85 and 0.82, and the RPD of 2.9 and 2.4, respectively. The ANN model for hardness of cooked GBR should be implemented to the practical use in GBR production factories for the quality assurance and further updating using more samples and several brands to obtain the robust models.

An approach to the implementation of neural network for real-time cryptographic data protection with symmetric keys oriented on embedded systems is presented. This approach is valuable especially for onboard communication systems in unmanned aerial vehicles (UAV) because of suitability to hardware implementation. In this study we evaluate the possibility to build such a system in hardware implementation at FPGA. The proposed integrated neural network approach for real-time cryptographic data protection was based on theoretical foundations of neural network cryptographic data protection, new algorithms and structures of neural network data encryption and decryption, modern hardware components with programmable structure. The development and implementation of the on-board neural network system for cryptographic data protection in real-time are based on the following principles: variable composition of equipment; modularity; conveyorization and spatial parallelism; hardware and software modification and making them suitable for data encryption and decryption. The tabular-algorithmic method for calculation of the scalar product has been improved that makes it possible to perform fast calculation of the scalar product of fixed-point and floating-point input data by bringing them to the largest common order of weights and forming tables of macro-partial products. Components of neural network cryptographic data encryption and decryption have been developed on the processor core supplemented by specialized scalar product calculation modules. The specialized hardware for neural network cryptographic data encryption was developed using VHDL for equipment programming in the Quartus II development environment ver. 13.1 and the appropriate libraries, and implemented on the basis of the FPGA EP3C16F484C6 Cyclone III family.

Geometrical assessments of human skull have been done based on the anatomical landmarks. Automatic detection of the landmarks, if developed, will be a great help not only medically, but also anthropologically. An automated system with multi-phased deep learning networks to predict three-dimensional coordinate values of cranio-facial landmarks, was developed. From a publicly available database, computed tomography images of craniofacial area were obtained. They were digitally reconstructed into three-dimensional models. Sixteen anatomical landmarks were plotted on each of the models and coordinate values of them were recorded. Three-phased regression deep learning networks were trained respectively with 90 training datasets. For evaluation, 30 testing datasets were employed. Three-dimensional error for the first phase, testing 30 data, was 11.60 pixels in average. (1 pixel = 500 / 512 mm) For the second phase, it was significantly improved to 4.66 pixels. For the third phase, it was significantly progressed to 2.88. This was comparable to the gaps between the landmarks, plotted by two experienced practitioners. Our proposing method of multi-phased prediction, coarse detection first and narrowing down the detection area, may be a possible solution, within the physical limitation of memory and computation.

Deepfakes are manipulated or altered images, or video, that are created using deep learning models with high levels of photorealism. The two popular methods of producing a deepfake are based on either convolutional neural networks (CNN), or autoencoders. Deepfakes created using CNN comparatively show higher qualities of realism, yet oftentimes leave artifacts and distortions in the generated media that can be detected using machine learning and deep learning algorithms. In recent years, there has been an influx of periocular image and video data because of the increase usage of face masks. By wearing masks, much of what is used for facial recognition is hidden, leaving only the periocular region visible to an observer. This loss of vital information leads to easier misidentification of media, allowing deepfakes to less likely be identified as fake. In this work, feature extraction methods, such as Scale-Invariant Feature Transform (SIFT), Histogram of Oriented Gradients (HOG), and CNN, are used to train an ensemble deep learning model to detect deepfakes in videos on a frame-by-frame level based on the periocular region. Our proposed model is able to distinguish original and manipulated images with accuracies around 98.9 percent, which is an improvement to previous works by combining SIFT and HOG for deepfake detection in convolutional neural networks.

Colonoscopy is used for colorectal cancer (CRC) screening. Extracting details of the colonoscopy findings from free text in electronic health records (EHRs) can be used to determine patient risk for CRC and colorectal screening strategies. In this study, we developed and evaluated the accuracy of a deep learning model framework to extract information for the clinical decision support system to analyze relevant free-text colonoscopy reports, including indications, pathology, and findings notes. The Bio-Bi-LSTM-CRF framework was developed using Bidirectional Long Short-term Memory (Bi-LSTM) and Conditional Random Fields (CRF) to extract several clinical features from these free-text reports, including indications for the colonoscopy, findings during the colonoscopy, and pathology of the resected material. We then trained the Bio-Bi-LSTM-CRF and existing Bi-LSTM-CRF models on 80% of 4,000 manually annotated notes obtained from the colonoscopy reports of 3,867 patients. The clinical notes were from a group of patients aged over 40 years old enrolled in four Veterans Affairs Medical Centers. A total of 10% of the remaining annotated notes were used to train hyperparameter, while the remaining 10% were used to evaluate the accuracy of our model (Bio-Bi-LSTM-CRF) and to compare the results with the outcomes obtained using Bi-LSTM-CRF. The results of our experiments showed that the bidirectional encoder representations by integrating dictionary function vector from Bio-Bi-LSTM-CRF and strategies character sequence embedding approach is an effective way to identify colonoscopy features from EHR-extracted clinical notes. Therefore, the Bio-Bi-LSTM-CRF model is concluded to be capable of creating new opportunities to identify patients at risk for colon cancer and to study their health outcomes.

A phenotype is the composite of an observable expression of a genome for traits in a given environment. The trajectories of phenotypes computed from an image sequence and timing of important events in a plant’s life cycle can be viewed as temporal phenotypes and indicative of the plant’s growth pattern and vigor. In this paper, we introduce a novel method called FlowerPhenoNet which uses deep neural networks for detecting flowers from multiview image sequences for high throughput temporal plant phenotyping analysis. Following flower detection, a set of novel flower-based phenotypes are computed, e.g., the day of emergence of the first flower in a plant’s life cycle, the total number of flowers present in the plant at a given time, the highest number of flowers bloomed in the plant, growth trajectory of a flower and the blooming trajectory of a plant. To develop a new algorithm and facilitate performance evaluation based on experimental analysis, a benchmark dataset is indispensable. Thus, we introduce a benchmark dataset called FlowerPheno which comprises image sequences of three flowering plant species, e.g., sunflower, coleus, and canna, captured by a visible light camera in a high throughput plant phenotyping platform from multiple view angles. The experimental analyses on the FlowerPheno dataset demonstrate the efficacy of the FlowerPhenoNet.

The environmental issues we are currently facing require long-term prospective efforts for sustainable growth. Renewable energy sources seem to be one of the most practical and efficient alternatives in this regard. Understanding a nation's pattern of energy use and renewable energy production is crucial for developing strategic plans. No previous study has been performed to explore the dynamics of power consumption with the change in renewable energy production on a country-wide scale. In contrast, a number of deep learning algorithms demonstrated acceptable performance while handling sequential data in the era of data-driven predictions. In this study, we developed a scheme to investigate and predict total power consumption and renewable energy production time series for eleven years of data using a Recurrent Neural Network (RNN). The dynamics of the interaction between the total annual power consumption and renewable energy production are investigated through extensive Exploratory Data Analysis (EDA) and a feature engineering framework. The performance of the model is found satisfactory through the comparison of the predicted data with the observed data, visualization of the distribution of the errors and Root Mean Squared Error (RMSE) value of 0.084. Higher performance is achieved through the increase in the number of epochs and hyperparameter tuning. The proposed framework can be used and transferred to investigate the trend of renewable energy production and power consumption and predict the future scenarios for different communities. Incorporation of the cloud-based platform into the proposed pipeline may lead to real-time forecasting.

Free-space communication is a leading component in global communications. Its advantages relate to a broader signal spread, no wiring, and ease of engagement. Satellite communication services became recently attractive to mega-companies that foresee an excellent opportunity to connect disconnected remote regions, serve emerging machine-to-machine communication, Internet-of-things connectivity, and more. Satellite communication links suffer from arbitrary weather phenomena such as clouds, rain, snow, fog, and dust. In addition, when signals approach the ground station, it has to overcome buildings blocking the direct access to the ground station. Therefore, satellites commonly use redundant signal strength to ensure constant and continuous signal transmission, resulting in excess energy consumption, challenging the limited power capacity generated by solar energy or the fixed amount of fuel. This research proposes LTSM, an artificial recurrent neural network technology that provides a time-dependent prediction of the expected attenuation level due to rain and fog and the signal strength that remained after crossing physical obstacles surrounding the ground station. The satellite transmitter is calibrated accordingly. The satellite outgoing signal strength is based on the predicted signal strength to ensure it will remain strong enough for the ground station to process it. The instant calibration eliminates the excess use of energy resulting in energy savings.

The emergence of Artificial Intelligence (AI) driven Keyword Spotting (KWS) technologies has revolutionized human to machine interaction. Yet, the challenge of end-to-end energy efficiency, memory footprint and system complexity of current Neural Network (NN) powered AI-KWS pipelines has remained ever present. This paper evaluates KWS utilizing a learning automata powered machine learning algorithm called the Tsetlin Machine (TM). Through significant reduction in parameter requirements and choosing logic over arithmetic based processing, the TM offers new opportunities for low-power KWS while maintaining high learning efficacy. In this paper we explore a TM based keyword spotting (KWS) pipeline to demonstrate low complexity with faster rate of convergence compared to NNs. Further, we investigate the scalability with increasing keywords and explore the potential for enabling low-power on-chip KWS.

Cold chain logistics of Agricultural Products demand forecasting can provide the scientific basis for the country to formulate logistics strategy, which further promotes the development of social economy and the improvement of living standards in China. In this paper, a new mathematical combined model is proposed to Agricultural Products Demand. Shandong, one of a China’s province, serves as the main producer and distributor of agricultural products. Based on the index system created from multiple related factors influencing cold chain logistics demand of agricultural products in Shandong, this paper employs principal component analysis to reduce the dimension of various indexes and predicts principal components with time series. Thereafter, multiple linear regression model and neural network model were constructed to forecast the cold chain logistics demand of agricultural products in Shandong, and their combined forecast models were compared. What's more, the paper provides insight for reference and decision-making concerning the development of cold chain logistics industry of agricultural products in Shandong province.

Reemergence of deep Neural Networks (CNNs) has lead to high-performance supervised learning algorithms for the Electro-Optical (EO) domain classification and detection problems. This success is possible because generating huge labeled datasets has become possible using modern crowdsourcing labeling platforms such as Amazon Mechanical Turk that recruit ordinary people to label data. Unlike the EO domain, labeling the Synthetic Aperture Radar (SAR) domain data can be a lot more challenging and for various reasons using crowdsourcing platforms is not feasible for labeling the SAR domain data. As a result, training deep networks using supervised learning is more challenging in the SAR domain. In the paper,we present a new framework to train a deep neural network for classifying Synthetic Aperture Radar (SAR) images by eliminating the need for huge labeled dataset. Our idea is based on transferring knowledge from a related EO domain problem, where labeled data is easy to obtain. We transfer knowledge from the EO domain through learning a shared invariant cross-domain embedding space that is also discriminative for classification. To this end, we train two deep encoders that are coupled through their last year to map data points from the EO and the SAR domains to the shared embedding space such that the distance between the distributions of the two domains is minimized in the latent embedding space. We use the Sliced Wasserstein Distance (SWD) to measure and minimize the distance between these two distributions and use a limited number of SAR label data points to match the distributions class-conditionally. As a result of this training procedure, a classifier trained from the embedding space to the label space using mostly the EO data would generalize well on the SAR domain. We provide theoretical analysis to demonstrate why our approach is effective and validate our algorithm on the problem of ship classification in the SAR domain by comparing against several other learning competing approaches.

Parmigiano Reggiano cheese is one of the most appreciated and consumed food worldwide, especially in Italy, for its high content of nutrients and for its taste. However, these characteristics make this product subject to counterfeiting in different forms. In this study, a novel method based on an electronic nose has been developed in order to investigate the potentiality of this tool to distinguish rind percentage in grated Parmigiano Reggiano packages that should be lower than 18%. Different samples in terms of percentage, seasoning and rind working process were considered to tackle the problem at 360°. In parallel, GC-MS technique was used to give a name to the compounds that characterize Parmigiano and to relate them with sensors responses. Data analysis consisted of two stages: multivariate analysis (PLS) and classification made in a hierarchical way with PLS-DA ad ANNs. Results are promising in terms of correct classification of the samples. The classification rate is higher for ANNs than PLS-DA, reaching also 100% values.