Search

• Five Machine Learning (ML) models compared for wave height time series prediction. • Complex ML models do not outperform simple AR in wave height time series prediction. • Comment to related papers: signal decomposition in test set series is WRONG. Significant Wave Height (SWH) is crucial in many aspect of ocean engineering. The accurate prediction of SWH has therefore been of immense practical value. Recently, Artificial Intelligence (AI) time series prediction methods have been widely used for single-point short-term SWH time-series forecasting, resulting in many AI-based models claiming to achieve good results. However, the extent to which these complex AI models can outperform traditional methods has largely been overlooked. This study compared five different models - AutoRegressive (AR), eXtreme Gradient Boosting (XGB), Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), and WaveNet - for their performance on SWH time series prediction at 16 buoy locations. Surprisingly, the results suggest that the differences of performance among different models are negligible, indicating that all these AI models have only "learned" the linear auto-regression from the data. Additionally, we noticed that many recent studies used signal decomposition method for such time series prediction, and most of them decomposed the test sets, which is WRONG. [ABSTRACT FROM AUTHOR]

Artificial Intelligence has found wide application, but also poses risks due to unintentional or malicious tampering during deployment. Regular checks are therefore necessary to detect and prevent such risks. Fragile watermarking is a technique used to identify tampering in AI models. However, previous methods have faced challenges including risks of omission, additional information transmission, and inability to locate tampering precisely. In this paper, we propose a method for detecting tampered parameters and bits, which can be used to detect, locate, and restore parameters that have been tampered with. We also propose an adaptive embedding method that maximizes information capacity while maintaining model accuracy. Our approach was tested on multiple neural networks subjected to attacks that modified weight parameters, and our results demonstrate that our method achieved great recovery performance when the modification rate was below 20%. Furthermore, for models where watermarking significantly affected accuracy, we utilized an adaptive bit technique to recover more than 15% of the accuracy loss of the model. • Fragile watermarking for rigorous neural network integrity detection. • Watermarking ensures 100% tampering detection, precise location, and recovery. • Watermarking fragility by combining Mutual- and Self-neural net parameters. • Incorporates watermarking with neural net feature, maintaining watermarked models' performance via gradient descent. • Bit-level training controls step size with carefully set hyperparameters. [ABSTRACT FROM AUTHOR]

• We analyzed older driver at-fault crashes at unsignalized intersections in Alabama. • We estimated random parameter models. • We introduced a deep learning approach based on Artificial Neural Networks. • We identified variables that increase the likelihood of injuries and fatalities. • We discussed possible countermeasures to address older driver at-fault crashes. Introduction: The research described in this paper explored the factors contributing to the injury severity resulting from the male and female older driver (65 years and older) at-fault crashes at unsignalized intersections in Alabama. Method: Random parameter logit models of injury severity were estimated. The estimated models identified a variety of statistically significant factors influencing the injury severities resulting from older driver at-fault crashes. Results : According to these models, some variables were found to be significant only in one model (male or female) but not in the other one. For example, variables such as driver under the influence of alcohol/drugs, horizontal curve, and stop sign were found significant only in the male model. On the other hand, variables such as intersection approaches on tangents with flat grade, and driver older than 75 years were found significant only in the female model. In addition, variables such as making turning maneuver, freeway-ramp junction, high speed approach, and so forth were found significant in both models. Estimation findings showed that two parameters in the male model and another two parameters in the female model could be modeled as random parameters, indicating their varying influences on the injury severity due to unobserved effects. In addition to the random parameter logit approach, a deep learning approach based on Artificial Neural Networks was introduced to predict the outcome of the crashes based on 164 variables that are listed in the crash database. The artificial intelligence (AI)-based method achieved an accuracy of 76% indicating the role of the variables in deciding the final outcome. Practical Applications: Future plans are set to study the use of AI on large sized datasets to achieve a relatively high-performance, and hence to be able to identify which variables contribute the most to the final outcome. [ABSTRACT FROM AUTHOR]

The increasing integration of distributed energy resources has led to heightened complexity in distribution network models, posing challenges of uncertainty and volatility to the operation and control of distribution networks. Simultaneously, the widespread deployment of metering and sensing devices has furnished data-driven approaches with copious data sources. As a potent tool for data mining, artificial intelligence (AI) algorithms have the capacity to harness extensive datasets, thus furnishing comprehensive energy services for distribution grids. This paper provides a comprehensive exposition of AI techniques applied in the realm of distribution network voltage control (DNVC). Initially, a brief overview is provided for the four voltage control frameworks and the data calibration of DNVC, serving as the foundation for various voltage control algorithms. Subsequently, in ascending order of algorithmic advancement, the application of improved heuristic algorithms, deep learning (DL) algorithms structured around artificial neural networks (ANNs), and deep reinforcement learning (DRL) algorithms in the DNVC domain are sequentially introduced, and the relationships among the three AI technologies in framework, control principles, and performance are systematically analyzed. Additionally, attention is directed toward the algorithmic progress made in addressing specific challenges within the DNVC process. Lastly, building upon extant research, the paper summarizes four prominent developmental directions for AI in DNVC: safety exploration, information privacy, multi-feeder collaboration, and heightened uncertainty. Corresponding solutions to each of these directions are individually discussed. • The research status and applications of mainly used AI methods in distribution network voltage control is reviewed. • The relationships among the three artificial intelligence methods are systematically analyzed. • The existing algorithms, along with their strengths, limitations, and relationshipsare highlighted. • The research directions of artificial intelligence methods in distribution network voltage control are discussed in detail. [ABSTRACT FROM AUTHOR]

To mitigate the high cost of integrated ship design, in this paper, we propose an innovative integrated hull design methodology utilizing artificial intelligence-aided design technology. The AIAD methodology leverages a deep neural network model constructed from multiple fidelities datasets to curtail the exponential growth in data requirements resulting from the proliferation of design variables caused by integration. Combining the proposed bump-free deformation method and optimization algorithms with the multi-fidelity deep network, the AIAD-based design approach facilitates the achievement of optimal hull design solutions with remarkable fidelity while minimizing design workload. Utilizing the developed design method, we performed a numerical test on integrated-hull design for container ships. Through a comparative analysis and sensitive analysis of design results and workload, the optimized results obtained demonstrated the effectiveness of the proposed AIAA integrated-hull design methodology in achieving optimal designs while preserving high fidelity and low workload. The proposed new approach has shown promises to expedite ship integrated-hull design, addressing the longstanding issue of its high cost. • An artificial intelligence-based design method is proposed, reducing the workload brought by the integrated-hull design. • Balanced the conflicting relationship between fidelity and cost of the deep learning model by using multi-fidelity datasets. • A bump-free deformation method is provided, which solves the local concave-convex after deformation. • Carried out the integrated-hull design of a container ship hull, achieving a 3% reduction in resistance. [ABSTRACT FROM AUTHOR]

Extensive research has been conducted on analyzing human movements, driven by its diverse practical applications such as human–robot interaction, human learning, and clinical diagnosis. However, the current state-of-the-art still encounters scientific challenges when it comes to modeling human movements. There are two key aspects that need to be addressed. Firstly, new models should consider the stochastic nature of human movement and the physical structure of the human body to accurately predict the patterns in full-body motion descriptors over time. Secondly, while deep learning algorithms have been utilized, they lack explainability in terms of predicting body posture sequences, making it essential to improve their comprehensible representation of human movement. This paper aims to tackle these challenges by presenting three innovative methods for creating explainable representations of human movement. The study formulates human body movement as a state-space model based on the Gesture Operational Model (GOM). Model parameters are estimated through either one-shot training employing Kalman Filters or data-intensive training utilizing artificial neural networks. The trained models are utilized for analyzing the dexterity of expert professionals in full-body movements, enabling the identification of dynamic associations between body joints and gesture recognition. Additionally, these models are employed to generate artificial professional movements. • Three innovative methods for explainable human movement models. • Body dexterity analysis: Models reveal how artisans and operators perform tasks. • Minimal motion descriptors to enhance task recognition. • Artificial generation of full-body professional movements. • Extensive experimentation with real-world datasets. [ABSTRACT FROM AUTHOR]

An in-depth understanding of uncertainty is the first step to making effective decisions under uncertainty. Machine/deep learning (ML/DL) has been hugely leveraged to solve complex problems involved with processing high-dimensional data. However, reasoning and quantifying different uncertainties to achieve effective decision-making have been much less explored in ML/DL than in other Artificial Intelligence (AI) domains. In particular, belief/evidence theories have been studied in Knowledge representation and reasoning (KRR) since the 1960s to reason and measure uncertainties to enhance decision-making effectiveness. Based on our in-depth literature review, only a few studies have leveraged mature uncertainty research in belief/evidence theories in ML/DL to tackle complex problems under different types of uncertainty. Our present survey paper discusses major belief theories and their core ideas dealing with uncertainty causes and types and quantifying them, along with the discussions of their applicability in ML/DL. Particularly, we discuss three main approaches leveraging belief theories in Deep Neural Networks (DNNs), including Evidential DNNs, Fuzzy DNNs, and Rough DNNs, in terms of their uncertainty causes, types, and quantification methods along with their applicability in diverse problem domains. Through an in-depth understanding of the extensive survey on this topic, we discuss insights, lessons learned, limitations of the current state-of-the-art bridging belief theories and ML/DL, and future research directions. This paper conducts an extensive survey by bridging belief theories and deep learning in reasoning and quantifying uncertainty to help researchers initiate uncertainty and decision-making research. • Identifying the causes and types of uncertainty in various belief models. • Uncertainty reasoning and quantification of belief models for decision-making. • Applying uncertainty quantification of belief models into deep learning. • Comparing uncertainty consideration in belief models and deep learning. • Insights, lessons learned, limitations, and future research directions. [ABSTRACT FROM AUTHOR]

Artificial intelligence, and in particular deep learning using convolutional neural networks, has been used extensively for image classification and segmentation, including on medical images for diagnosis and prognosis prediction. Use in radiotherapy prognostic modelling is still limited, however, especially as applied to toxicity and tumour response prediction from radiation dose distributions. We review and summarise studies that applied deep learning to radiotherapy dose data, in particular studies that utilised full three-dimensional dose distributions. Ten papers have reported on deep learning models for outcome prediction utilising spatial dose information, whereas four studies used reduced dimensionality (dose volume histogram) information for prediction. Many of these studies suffer from the same issues that plagued early normal tissue complication probability modelling, including small, single-institutional patient cohorts, lack of external validation, poor data and model reporting, use of late toxicity data without taking time-to-event into account, and nearly exclusive focus on clinician-reported complications. They demonstrate, however, how radiation dose, imaging and clinical data may be technically integrated in convolutional neural networks-based models; and some studies explore how deep learning may help better understand spatial variation in radiosensitivity. In general, there are a number of issues specific to the intersection of radiotherapy outcome modelling and deep learning, for example translation of model developments into treatment plan optimisation, which will require further combined effort from the radiation oncology and artificial intelligence communities. • The use of deep learning for radiotherapy outcome prediction is reviewed. • Ten studies used convolutional neural networks and spatial dose for prediction. • Studies were generally small and single institutional, with lack of external validation. • Deep learning may explore spatial variation in radiosensitivity, but methodology underdeveloped. [ABSTRACT FROM AUTHOR]

Online social networks can be used for mental healthcare monitoring using Artificial Intelligence and Machine Learning techniques for detecting various mental health disorders and corresponding risk assessment. Recent research in this domain has primarily been focused on leveraging deep neural networks and various Transformer based Large Language Models, which have now become state-of-the-art for most natural language processing and computational linguistic tasks due to their unmatched prediction accuracy. Unlike conventional machine learning algorithms, these deep neural networks are black box architectures, where it is difficult to interpret and explain their predicted outcome. However, a black box classification outcome is insufficient for healthcare applications. Such systems will not be widely adopted and trusted by healthcare practitioners if they are not able to understand and explain the reasoning behind the predicted decisions made by an AI and ML based healthcare diagnostic system. The key objective of our research is to demonstrate the applications of model agnostic, post hoc surrogate XAI techniques for providing explainability to classification decisions of pretrained LLMs (Transformers) based mental healthcare diagnostic systems fine-tuned (or trained) to detect depressive and suicidal behavior using UGC from online social networks. For this, we have used the two most recent and popular techniques, SHAP and LIME. We have conducted extensive and in-depth experiments with four datasets and six pretrained LLMs, three of which have already been domain-adapted using mental health related datasets. We have also performed Few Shot Learning experiments with these three pretrained mental health domain-adapted LLMs. The results of qualitative and descriptive data analysis in this paper demonstrate that in order to build a comprehensive understanding of a person's psychological state, emotion, and behavior and to discover the causes, symptoms, and triggers of mental health issues, it is essential to utilize e X pl AI nable (XAI) techniques with Transformer based LLMs (supervised). Alternatively, Transformer based unsupervised topic modeling technique BERTopic may be used for mental health risk monitoring and cause or symptom extraction when supervised training of LLMs is not feasible due to dataset annotation or availability challenges. [ABSTRACT FROM AUTHOR]

Robustness has become an important consideration in deep learning. With the help of explainable AI, mismatches between an explained model's decision strategy and the user's domain knowledge (e.g. Clever Hans effects) have been identified as a starting point for improving faulty models. However, it is less clear what to do when the user and the explanation agree. In this paper, we demonstrate that acceptance of explanations by the user is not a guarantee for a machine learning model to be robust against Clever Hans effects, which may remain undetected. Such hidden flaws of the model can nevertheless be mitigated, and we demonstrate this by contributing a new method, Explanation-Guided Exposure Minimization (EGEM), that preemptively prunes variations in the ML model that have not been the subject of positive explanation feedback. Experiments demonstrate that our approach leads to models that strongly reduce their reliance on hidden Clever Hans strategies, and consequently achieve higher accuracy on new data. • Explainable AI may not detect all Clever Hans effects in practical ML scenarios. • Hidden Clever Hans effects can be mitigated using Exposure Minimization. • We propose a solution in the form of a soft-pruning scheme on network activations. • We demonstrate the practical benefits of our approach on image and text data. [ABSTRACT FROM AUTHOR]

Nowadays, Artificial intelligence (AI), combined with the digitalization of healthcare, can lead to substantial improvements in Patient Care, Disease Management, Hospital Administration, and supply chain effectiveness. Among predictive analytics tools, time series forecasting represents a central task to support healthcare management in terms of bookings and medical services predictions. In this context, the development of flexible frameworks to provide robust and reliable predictions became a central point in this healthcare innovation process. This paper presents and discusses a multi-source time series fusion and forecasting framework relying on Deep Learning. By combining weather, air-quality and medical bookings time series through a feature compression stage which preserves temporal patterns, the prediction is provided through a flexible ensemble technique based on machine learning models and a hybrid neural network. The proposed system is able to predict the number of bookings related to a specific medical examination for a 7-days horizon period. To assess the proposed approach's effectiveness, we rely on time series extracted from a real dataset of administrative e-health records provided by the Campania Region health department, in Italy. • We propose a multi-source time-series fusion methodology relying on Deep Learning. • We design a robust framework for daily 7-days-horizon predictions. • We assess the proposed approach on a real medical bookings dataset. [ABSTRACT FROM AUTHOR]

Gastrointestinal cancers account for approximately 20% of all cancer diagnoses and are responsible for 22.5% of cancer deaths worldwide. Artificial intelligence–based diagnostic support systems, in particular convolutional neural network (CNN)–based image analysis tools, have shown great potential in medical computer vision. In this systematic review, we summarise recent studies reporting CNN-based approaches for digital biomarkers for characterization and prognostication of gastrointestinal cancer pathology. Pubmed and Medline were screened for peer-reviewed papers dealing with CNN-based gastrointestinal cancer analyses from histological slides, published between 2015 and 2020.Seven hundred and ninety titles and abstracts were screened, and 58 full-text articles were assessed for eligibility. Sixteen publications fulfilled our inclusion criteria dealing with tumor or precursor lesion characterization or prognostic and predictive biomarkers: 14 studies on colorectal or rectal cancer, three studies on gastric cancer and none on esophageal cancer. These studies were categorised according to their end-points: polyp characterization, tumor characterization and patient outcome. Regarding the translation into clinical practice, we identified several studies demonstrating generalization of the classifier with external tests and comparisons with pathologists, but none presenting clinical implementation. Results of recent studies on CNN-based image analysis in gastrointestinal cancer pathology are promising, but studies were conducted in observational and retrospective settings. Large-scale trials are needed to assess performance and predict clinical usefulness. Furthermore, large-scale trials are required for approval of CNN-based prediction models as medical devices. • Computational medical image analysis is rising in gastrointestinal cancer pathology. • Several studies deal with CNN–based identification of biomarkers on HE slides. • Results are promising but studies are at early stages. • Further studies are needed to assess performance and potential clinical usefulness. [ABSTRACT FROM AUTHOR]

Early fault detection, both in equipment and the products in process, is of paramount importance in industrial processes to ensure the quality of the final product, avoid abnormal operating conditions, expensive repairs, and even production process shutdown. The growing complexity of industrial systems and the increase in the amount of available data have encouraged the development of intelligent systems for automatic fault prediction/detection, mainly based on Industry 4.0 technologies and, particularly, those based on deep learning methodologies. However, the vast majority of proposals and research carried out to date define specific solutions for specific cases, which still requires a high level of expert knowledge for scaling the solutions to industrial environments. Actually, one of the major issues towards the industrialization of deep learning solutions is the determination of the optimal, or near-optimal, hyper-parameters. In this paper, we propose a low-level set-up effort intelligent failure detection system that integrates deep neural networks with a Bayesian Optimization algorithm for self-tuning of the system hyper-parameters. In addition, to facilitate the industrialization of the proposal and its incorporation into current industrial systems, we embedded the proposal in our previously formulated and tested Simulai architecture which allows for containing and interaction with multiple and heterogeneous technological components of manufacturing processes. Finally, our proposal is tested in two real cases of a different nature. The obtained results show a successful performance and demonstrate the easy online integration and interaction in a real production system. • Industrialization of artificial intelligence through a self-adjusting system. • Improvement of the decision-making process in the detection of industrial failures. • Optimization of deep learning for the improvement of industrial health and quality. • Demonstration of the proposal's adaptability to real cases of different natures. [ABSTRACT FROM AUTHOR]

Modern artificial intelligence systems rely heavily on deep learning techniques. However, deep neural networks are easily disturbed by adversarial objects. Adversarial examples are beneficial to improve the robustness of the 3D neural network model and enhance the stability of the artificial intelligence system. At present, most 3D adversarial attack methods perturb the entire point cloud to generate adversarial examples, which results in high perturbation costs and low operability in the physical world. In this paper, we propose an adaptive local adversarial attack method (AL-Adv) on 3D point clouds to generate adversarial point clouds. First, we analyze the vulnerability of the 3D network model and extract the salient regions of the input point cloud, namely the vulnerable regions. Second, we propose an adaptive gradient attack algorithm that targets salient regions. The proposed attack algorithm adaptively assigns different disturbances in different directions of the three-dimensional coordinates of the point cloud. Experimental results show that our proposed AL-Adv method achieves a higher attack success rate than the global attack method. Specifically, the adversarial examples generated by AL-Adv demonstrate good imperceptibility and small generation costs. • The proposed AL-Adv method focuses on local regions of point clouds. • We design a novel adaptive gradient attack algorithm for local regions. • The proposed method AL-Adv achieves a higher attack success rate. • The proposed AL-Adv method generates adversarial examples with good imperceptibility. • The proposed AL-Adv method outperforms the global attack method. [ABSTRACT FROM AUTHOR]

[Display omitted] • We developed AICropCAM, an edge-computing enabled camera system for crop monitoring. • It integrated Raspberry Pi and Arduino for image processing and LoRa communication. • It ran a stack of four deep neural networks to characterize multiple plant/canopy parameters. • We quantified the power consumption and speed of the system for edge image-processing. • AICropCAM is a next-generation enabling technology for real-time, low-latency Ag applications. Precision Agriculture (PA) promises to meet the future demands for food, feed, fiber, and fuel while keeping their production sustainable and environmentally friendly. PA relies heavily on sensing technologies to inform site-specific decision supports for planting, irrigation, fertilization, spraying, and harvesting. Traditional point-based sensors enjoy small data sizes but are limited in their capacity to measure plant and canopy parameters. On the other hand, imaging sensors can be powerful in measuring a wide range of these parameters, especially when coupled with Artificial Intelligence. The challenge, however, is the lack of computing, electric power, and connectivity infrastructure in agricultural fields, preventing the full utilization of imaging sensors. This paper reported AICropCAM, a field-deployable imaging framework that integrated edge image processing, Internet of Things (IoT), and LoRaWAN for low-power, long-range communication. The core component of AICropCAM is a stack of four Deep Convolutional Neural Networks (DCNN) models running sequentially: CropClassiNet for crop type classification, CanopySegNet for canopy cover quantification, PlantCountNet for plant and weed counting, and InsectNet for insect identification. These DCNN models were trained and tested with >43,000 field crop images collected offline. AICropCAM was embodied on a distributed wireless sensor network with its sensor node consisting of an RGB camera for image acquisition, a Raspberry Pi 4B single-board computer for edge image processing, and an Arduino MKR1310 for LoRa communication and power management. Our testing showed that the time to run the DCNN models ranged from 0.20 s for InsectNet to 20.20 s for CanopySegNet, and power consumption ranged from 3.68 W for InsectNet to 5.83 W for CanopySegNet. The classification model CropClassiNet reported 94.5 % accuracy, and the segmentation model CanopySegNet reported 92.83 % accuracy. The two object detection models PlantCountNet and InsectNet reported mean average precision of 0.69 and 0.02 for the test images. Predictions from the DCNN models were transmitted to the ThingSpeak IoT platform for visualization and analytics. We concluded that AICropCAM successfully implemented image processing on the edge, drastically reduced the amount of data being transmitted, and could satisfy the real-time need for decision-making in PA. AICropCAM can be deployed on moving platforms such as center pivots or drones to increase its spatial coverage and resolution to support crop monitoring and field operations. [ABSTRACT FROM AUTHOR]

Predicting electricity demand (G) is crucial for electricity grid operation and management. In order to make reliable predictions, model inputs must be analyzed for predictive features before they can be incorporated into a forecast model. In this study, a hybrid multi-algorithm framework is developed by incorporating Artificial Neural Networks (ANN), Encoder-Decoder Based Long Short-Term Memory (EDLSTM) and Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICMD). Following the partitioning of data, the G time-series are decomposed into multiple time-series using the ICEEMDAN algorithm, with partial autocorrelation applied to training sets to determine lagged features. We combine lagged inputs into a predictive framework where G components with the highest frequency are predicted with an ANN model, while remaining components are predicted with an EDLSTM model. To generate the results, all IMF components' predictions are merged using ICMD-ANN-EDLSTM hybrid models. A comparison is made between this objective model and standalone models (ANN, RFR, LSTM), hybrid models (CLSTM), and three decomposition-based hybrid models. Based on the results, the Relative Mean Absolute Error at Duffield Road substation was ≈ 2. 82 % , ≈ 4. 15 % , ≈ 3. 17 % , ≈ 6. 41 % , ≈ 6. 60 % , ≈ 6. 49 % , and ≈ 6. 602 % , compared to ICMD-RFR-LSTM, ICMD-RFR-CLSTM, LSTM, CLSTM, RFR, and ANN. According to statistical score metrics, the hybrid ICMD-ANN-EDLSTM model performed better than other benchmark models. Further, the results show that the hybrid ICMD-ANN-EDLSTM model can not only detect seasonality in G data, but also predict and analyze electricity market demand to add valuable insight to market analysis. [Display omitted] • Predicting electricity demand is challenging due to non-linear and non-stationary features. • A hybrid deep-leaning based multi-method approach is proposed in this paper. • Data decomposition-based encoder–decoder neural networks algorithm is proposed. • The algorithm adopts a neural network for high frequency and a deep LSTM model for low frequency input data. [ABSTRACT FROM AUTHOR]

Recently, facial expression recognition (FER) methods have achieved significant progress. However, FER is still challenged by factors such as uneven illumination and low-quality expression images. Exploring the potential of facial expression features to achieve robust FER is particularly important. Inspired by the Transformer's excellent performance in modeling long-range dependencies and in vision tasks, this paper proposes a Transformer Block Enhancement Module (TBEM) for enhancing the feature representation of facial expressions. The proposed module contains a Channel Enhancement (CE) block and a Spatial Enhancement (SE) block. The CE block can adaptively enhance the expression features on the channel dimension by effectively leveraging the channel dependency information, while the SE block enhances the expression features on the spatial dimension by integrating the spatial dependency information. TBEM can output a more robust expression representation by combining CE and SE. The proposed artificial intelligence learning module greatly improves the recognition accuracy of FER engineering tasks. To further illustrate the application of TBEM in real-world FER engineering, three engineering problems are used for verification. Extensive experiments demonstrate that the proposed method improves FER performance by focusing on more accurate decisional features and can be easily embedded into regular convolutional neural network models to help them improve the accuracy on FER tasks by about 2.64%–3.03%. The results show the proposed method achieves accuracies of 90.57% on FERPlus, 89.41% on RAFDB basic and 68.43% on RAFDB compound, respectively. The proposed method also provides a meaningful reference for further research on applying Transformer to other tasks. [ABSTRACT FROM AUTHOR]

• Data-driven methods offer excellent performance for process monitoring. • Their black-box nature without explanations, however, has limited their industrial adoption. • We propose a method for explaining the predictions by deep neural networks using alarm limits. • Two case studies demonstrate its utility to the plant operator as well as to the model developer. Deep Learning (DL) models are becoming the preferred approach for process monitoring due to their higher prediction accuracy; however, they are still viewed as black boxes. Explainable Artificial Intelligence (XAI) methods seek to address this shortcoming by providing explanations that are either global (explains the entire DL model) or local (explains the result for each individual sample). Due to nonlinearities and other complexities inherent in chemical processes, providing a local explanation is more suitable. This paper proposes a local XAI method that explains process monitoring results from a deep neural network (DNN) based on process alarms. The effectiveness of the proposed method is demonstrated using a CSTR case study and the Tennessee-Eastman benchmark process. Our results show that the explanations proffered by the proposed method assist operators in understanding DNN's prediction during online process monitoring. Additionally, during model development, it can offer insights that enable improvement of the DNN's architecture. [ABSTRACT FROM AUTHOR]

In the education industry, the transformation of educational concepts, the reform of teaching courses and the reorganization of teaching materials are the only way to strengthen quality education. Local music(LM) education is no exception. The development of LM teaching materials (TMs) suitable for various regions and schools is of great significance to deepen students' understanding of TMC and improve their interest in LM courses. Deep learning is a new artificial intelligence method that has been widely concerned and widely used in recent years. It is the hottest development content in artificial intelligence and is used in many fields. Deep neural network models have strong expressiveness, and model training is closer to the way the brain works. Therefore, based on the concept of deep learning, this paper builds a development system for local music teaching materials. Each school can refer to this system to develop textbooks, and in the process of TD, intelligent development software can be used to write textbooks. This paper analyzes the current situation of the development of TM teaching materials in 15 schools in a province through the methods of investigation, interview and data statistics, and expounds the attitudes and factors affecting the development of TMs from the perspectives of schools, teachers and students(TAS). The results show that there are 9 schools. Support the development of TMs. A total of 58.65 % of teachers and students are in favor of the development of TMs, but 39.29 % of teachers believe that their ability to develop traditional music TMs is average, and students do not have a deep understanding of LM. Therefore, the development of traditional music teaching It is necessary to conduct an in-depth discussion on the two teaching subjects of TAS. [ABSTRACT FROM AUTHOR]

With the development of artificial intelligence, the most representative deep learning has been applied to various fields, which is greatly influencing human society. However, deep neural networks (DNNs) are still a black-box model, and the process how they make decisions internally is still difficult to understand and control. At the same time, DNNs take up more hardware resources, resulting in high energy consumption. Therefore, it is significant to study the characteristics of deep AI models and deeply understand the interactions between parameters within AI models so as to improve the interpretability of DNNs, optimize their structure and increase their computational efficiency. In this paper, we propose a graph-based interpretability method for deep neural networks (GIMDNN). The running parameters of DNNs are modeled as a graph by using a kernel function or the Graph Transformer Networks (GTN), where the nodes of the graph are obtained by dimensional mapping of the parameters of the DNNs, and the edges are calculated by the Gaussian kernel function. The generated graphs are classified by a graph convolutional network (GCN). The association relationship between the adjacent layers and the running mechanism of DNNs are analyzed, and the importance of the parameters of each layer in the DNNs for the final classification result can be obtained. Convolutional neural networks (CNNs) are one of the most representative models in DNNs. The proposed method is experimentally evaluated on the CNNs. The experimental results show that the proposed method can interpret the associations among the weight parameters as well as the correlation between two adjacent layers. Therefore, the DNNs for special tasks, such as portable applications, edge computing, and so on, can be customized, the number of parameters can be reduced. It is valuable to interpret the operation and principle of CNNs. [ABSTRACT FROM AUTHOR]

Machine learning (ML) and deep learning (DL) are both branches of AI. As a form of AI, ML automatically adapts to changing datasets with minimal human interference. Deep learning is a subset of ML that uses artificial neural networks to imitate the learning process of the human brain. The 'black box' nature of ML and DL makes their inner workings difficult to understand and interpret. Deploying explainable artificial intelligence (XAI) can help explain why and how the output of ML and DL models are generated. As a result, understanding a model's functioning, behavior, and outputs can be garnered, reducing bias and error and improving confidence in decision-making. Despite providing an improved understanding of model outputs, XAI has received limited attention in construction. This paper presents a narrative review of XAI and a taxonomy of precepts and models to raise awareness about its potential opportunities for use in construction. It is envisaged that the opportunities suggested can stimulate new lines of inquiry to help alleviate the prevailing skepticism and hesitancy toward AI adoption and integration in construction. [ABSTRACT FROM AUTHOR]

Malignancy is the tendency of a medical condition to progressively become fatal and generally refers to the presence of cancerous cells that can spread, invade, and destroy tissues. In the United States, a new cancer diagnosis is given every 30 s, and the rates of diagnosis are not indicating any slowdown. The number of cancer cases is projected to increase by 35% over the next decade. Increasing cancer cases have attracted researchers to find new approaches for their identification and characterization. Skin diseases are a significant healthcare burden, affecting millions of people globally. The early detection and diagnosis of skin diseases are essential to prevent adverse health consequences, including the development of malignant skin tumors. Dermoscopy-based diagnosis of skin diseases is a time-consuming process, particularly in resource-limited healthcare settings. Recent advancements in computer vision algorithms have led to the development of several smartphone-based applications for early skin disease prognosis. In this paper, we have utilized a set of 16 distinctive convolutional neural network models, based on deep learning, to process over 45,000 high-quality images from the HAM10000 dataset, encompassing seven categories of skin diseases. Our models were able to effectively identify and classify skin diseases using lesion images as input. Dermoscopy-based diagnosis of skin diseases is a time-consuming process, particularly in resource-limited healthcare settings. Our proposed application for early skin disease detection aims to prevent unintended health consequences, which can result in wastage of operating hours, resources, and, in some cases, compromised patient health. We have benchmarked our models against one another to evaluate their effectiveness and efficiency, using performance measures such as accuracy, precision, sensitivity, specificity, F-score, and G-Mean. Our results demonstrate that most deployed models achieve an accuracy of up to 99%. [ABSTRACT FROM AUTHOR]

Several sectors have increased the use of lattice structures due its great capacity to reduce the structural mass with reasonable rigidity loss. However, these structures can present complex mechanical behaviors, impossible to be determined analytically. Surrogate models obtained from design of experiments have been shown to be promising but insufficient. Numerical models using finite elements are computationally expensive when optimization model updating is considered. In order to solve these problems, in this paper a deep learning model is trained in order to predict 16 different structural responses (static, dynamic and stability) of a complex composite isogrid structure tube. All input data were generated from a numerical finite element model. The results demonstrated substantial capacity of the deep learning model to fit the isogrid physical behavior and also predict the structural responses. In addition, it was also to predict the design parameters of the isogrid structure from the desired results, also demonstrating the good performance (R2 > 90 %) of the artificial intelligence model to perform this prediction. [ABSTRACT FROM AUTHOR]

• The layer sustainability analysis (LSA) framework is introduced to evaluate the behavior of layer-level representations of DNNs in dealing with network input perturbations using Lipschitz theoretical concepts. • A layer-wise regularized adversarial training (AT-LR) approach significantly improves the generalization and robustness of different deep neural network architectures for significant perturbations while reducing layer-level vulnerabilities. • AT-LR loss landscapes for each LSA MVL proposal can interpret layer importance for different layers, which is an intriguing aspect. Deep neural network models are used today in various applications of artificial intelligence, the strengthening of which, in the face of adversarial attacks is of particular importance. An appropriate solution to adversarial attacks is adversarial training, which reaches a trade-off between robustness and generalization. This paper introduces a novel framework (Layer Sustainability Analysis (LSA)) for the analysis of layer vulnerability in an arbitrary neural network in the scenario of adversarial attacks. LSA can be a helpful toolkit to assess deep neural networks and to extend the adversarial training approaches towards improving the sustainability of model layers via layer monitoring and analysis. The LSA framework identifies a list of Most Vulnerable Layers (MVL list) of the given network. The relative error, as a comparison measure, is used to evaluate representation sustainability of each layer against adversarial inputs. The proposed approach for obtaining robust neural networks to fend off adversarial attacks is based on a layer-wise regularization (LR) over LSA proposal(s) for adversarial training (AT). This means that the AT-LR procedure could be used with any benchmark adversarial attack to reduce the vulnerability of network layers and to improve conventional adversarial training approaches. The proposed idea performs well theoretically and experimentally for state-of-the-art multilayer perceptron and convolutional neural network architectures. Additionally, a measure named robustness and generalization score or R&G score is defined to better evaluate each adversarially trained model over a variety of significant perturbations. Compared with the AT-LR and its corresponding base adversarial training, the R&G score on Moon, MNIST, and CIFAR-10 benchmark datasets was increased by 56.52%, 75.82%, and 6.54%, respectively for more significant perturbations. The LSA framework is available and published at https://github.com/khalooei/LSA. [ABSTRACT FROM AUTHOR]

• Methodology and model for creating AI expert learning via AR and humans. • Positive climate impact when some services can be carried out remotely over AR calls. • Energy savings and CO₂ emissions reduction have been estimated. • Supporting the development of smart industries and autonomous operations. • Economic benefits as for health improvement and increased business productivity. Initially, Artificial Intelligence (AI) focused on diagnostics during the 70s and 80s. Unfortunately, it did not gain trust and few industries embraced it, mostly due to the extensive manual programming effort that AI required for interpreting data and act. In addition, the computer capacity, for handling the amounts of data necessary to train AI, was lacking the disc dimensions we are used to today, which made it go slowly. Not until the 2000 s confidence in AI was established in parallel with the introduction of new tools that was paving the way for PLS, PCA, ANN and soft sensors. Year 2011, IBM Watson (an AI application) was developed and won over the jeopardy champion. Today's machine learning (ML) such as "deep learning" and artificial neural networks (ANN) have created interesting use cases. AI has therefore regained confidence and industries are beginning to embrace where they see appropriate uses. Simultaneously, Internet of Things (IoT) tools have been introduced and made it possible to develop new capabilities such as virtual reality (VR), augmented reality (AR), mixed reality (MR) and extended reality (XR). These technologies are maturing and could be used in several application areas for the industries and form part of their digitalization journey. Furthermore, it is not only the industries that could benefit from introducing these technologies. Studies also show several areas and use cases where augmented reality has a positive impact, such as on students' learning ability. Yet few teachers know or use this technology. This paper evaluates and analyze AR, remote assistance tool for industrial purposes. The potential of the tool is discussed for frequent maintenance cases in the mining industry. Further on, if we look into the future, it is not surprising if we will be able to see that today's concepts of reality tools have evolved to become smarter by being trained by multimedia recognition and from people who have thus created an AI expert. Where the AI expert will support customers and be able to solve simple errors but also those that occur rarely and thus be a natural part of the solution for future completely autonomous processes for the industry. The article demonstrates a framework for creating smarter tools by combining AR, ML and AI and forms part of the basis creating the smarter industry of the future. Natural Language Processing (NLP) toolbox has been utilized to train and test an AI expert to give suitable resolutions to a specific maintenance request. The motivation for AR is the possible energy savings and reduction of CO₂ emissions in the maintenance field for all business trips that can be avoided. At the same time saving money for the industries and expert manhours that are spent on traveling and finally enhancing the productivity for the industries. Tests cases have verified that with AR, the resolution time could be significantly reduced, minimizing production stoppages by more than 50% of the time, which ultimately has a positive effect on a country's GDP. How much energy can be saved is predicted by the fact that 50% of all the world's business flights are replaced by one of the reality concepts and are estimated to amount to at least 50 Mton CO₂ per year. This figure is probably slightly higher as business trips also take place by other means of transport such as trains, buses, and cars. With today's volatile employees changing jobs more frequently, industry experts are becoming fewer and fewer. Since new employee stays for a maximum of 3–5 years per workplace, they will not stay long enough to become experts. Introducing an AI expert trained by today's experts, there is a chance that this knowledge can be maintained. [ABSTRACT FROM AUTHOR]

Multiphase flowmeters are important to monitor oil wells, as they allow operators to obtain a real-time estimate of the production. However, due to its high installation cost, uncertainty, and reading errors, the oil and gas industry started to invest in virtual meters, mathematical models that can obtain the oil and gas flow forecast using the instrumentation already available in the well, such as pressure and temperature sensors. The purpose of this article is to develop a virtual multiphase flow predictor using artificial intelligence models such as Long Short-Term Memory (LSTM), Gate Recurrent Units (GRU), Multi-Layer Perceptron (MLP) neural networks and deep learning. As input data, this paper uses a public database available on the Oil and Gas Authority (UK) to train the models and demonstrate the possibility of predicting multiphase flow with a reasonable error margin, proving that the method is efficient in estimating well production rates. • In the oil industry, it is important to measure the production from each well. • A Virtual Meter is a low-cost alternative. It can measure multiphase flow without any additional equipment. • This paper proposes a Virtual Meter using neural networks such as MLP, LSTM, GRU and deep learning combinations. • Prediction performance was evaluated using the Oil and Gas Authority (OGA) public database and it shows promising results. • Experiments using a combination of multiple layers shows this type of prediction takes advantage of deep learning techniques. [ABSTRACT FROM AUTHOR]

How to design System of Systems has been widely concerned in recent years, especially in military applications. This problem is also known as SoS architecting, which can be boiled down to two subproblems: selecting a number of systems from a set of candidates and specifying the tasks to be completed for each selected system. Essentially, such a problem can be reduced to a combinatorial optimization problem. Traditional exact solvers such as branch-bound algorithm are not efficient enough to deal with large scale cases. Heuristic algorithms are more scalable, but if input changes, these algorithms have to restart the searching process. Re-searching process may take a long time and interfere with the mission achievement of SoS in highly dynamic scenarios, e.g., in the Mosaic Warfare. In this paper, we combine artificial intelligence with SoS architecting and propose a deep reinforcement learning approach DRL-SoSDP for SoS design. Deep neural networks and actor–critic algorithms are used to find the optimal solution with constraints. Evaluation results show that the proposed approach is superior to heuristic algorithms in both solution quality and computation time, especially in large scale cases. DRL-SoSDP can find great solutions in a near real-time manner, showing great potential for cases that require an instant reply. DRL-SoSDP also shows good generalization ability and can find better results than heuristic algorithms even when the scale of SoS is much larger than that in training data. [ABSTRACT FROM AUTHOR]

The artificial intelligence (AI) revolution offers significant opportunities to capitalise on the growth of digitalisation and has the potential to enable the 'system of systems' approach required in increasingly complex infrastructure systems. This paper reviews the extent to which research in economic infrastructure sectors has engaged with fields of AI, to investigate the specific AI methods chosen and the purposes to which they have been applied both within and across sectors. Machine learning is found to dominate the research in this field, with methods such as artificial neural networks, support vector machines, and random forests among the most popular. The automated reasoning technique of fuzzy logic has also seen widespread use, due to its ability to incorporate uncertainties in input variables. Across the infrastructure sectors of energy, water and wastewater, transport, and telecommunications, the main purposes to which AI has been applied are network provision, forecasting, routing, maintenance and security, and network quality management. The data-driven nature of AI offers significant flexibility, and work has been conducted across a range of network sizes and at different temporal and geographic scales. However, there remains a lack of integration of planning and policy concerns, such as stakeholder engagement and quantitative feasibility assessment, and the majority of research focuses on a specific type of infrastructure, with an absence of work beyond individual economic sectors. To enable solutions to be implemented into real-world infrastructure systems, research will need to move away from a siloed perspective and adopt a more interdisciplinary perspective that considers the increasing interconnectedness of these systems. • Literature on the use of AI in infrastructure sectors is reviewed. • Types of AI, trends across sectors and purposes within infrastructure are identified. • Findings have been evaluated using an analysis framework. • The future of AI and the 'Survey PaperInternet of Things' in infrastructure are discussed. [ABSTRACT FROM AUTHOR]

Rockfall risk management has traditionally been built upon scenarios based on statistical analysis of historic frequency-magnitude rockfall patterns which allows for an estimation of the return period of rockfalls within given ranges of magnitude. Τhe availability of rockfall databases is thereby of crucial importance to quantify rockfall activity. In recent years, engineering geologists have been able to leverage light detection and ranging (LiDAR) technology with automated change detection and clustering workflows, for spatially accurate rockfall monitoring and inventorization. This wealth of digital data generates new research prospects for the development of an artificial intelligence (AI) system able to analyze change detection sequences and map spatio-temporal rockfall patterns by precisely capturing the rock slope evolution in near-real-time. However, prerequisite to such an advancement is the need to efficiently automate the process of classifying surface changes corresponding to rockfall. This task typically requires intensive human effort and is prone to multiple sources of human error like subjectivity, expertise, and experience that may differ from person to person involved in a particular project. This paper investigates the potential of integrating sophisticated deep learning architectures into dynamic rockfall database population processes, with the goal of relieving experts from the daunting task of manually classifying rockfall events within large loads of change detection data. Deep neural networks based on the pioneering PointNet and PointNet++ architectures for 3D point cloud learning are developed for this purpose based on a 5-year change detection database consisting of >8000 rock slope clusters of identified change for training, with scanning intervals ranging from 5 days to 6 months. The models are tested on the 536 clusters from the two last data acquisitions to simulate the real monitoring situation and subsequently on the most frequent of the campaigns to increase the probability of working with single-event clusters. The best-performing model achieves an accuracy of about 89% and 84% on the last and shortest campaign, respectively. The optimized deep learning models are further evaluated on a geologically different rockfall database achieving almost 93% accuracy in a location where discrete geomorphologic features such as steep rock outcrops and erosion channels are present. The study shows that although it is challenging to achieve generalization in rockfall detection, site-specific training of the proposed deep learning architecture can lead to high-level performance and support further advancements in rockfall risk management. • LiDAR monitoring provides detailed 3D point cloud models of rock slope deformations. • High-level feature learning from point clouds is achieved with modern deep learning. • Rockfall detection can be approached as a binary point cloud classification problem. • Global shape feature learning is suggested compared to local texture features. • The prediction confidence increases with the cluster volume. [ABSTRACT FROM AUTHOR]

[Display omitted] • Recent advances and future challenges and trends in flotation froth image analysis based on deep learning are reviewed. • Most advances are driven by transfer learning with publicly available neural networks pretrained on the ImageNet database. • Labelling of images poses a fundamental challenge to the application of these systems as soft sensors. • This can be ameliorated by use of methods, such as semi-supervised regression and k-shot learning. • The availability of large, curated froth image databases could serve to drive further development. Machine vision is widely used in the monitoring of froth flotation plants as a means to assist control operators on the plant. While these systems have a mature ability to analyse physical froth features, such as the colour of the froth and bubble size distributions, research has continued to focus on their use in automated control systems, which is not well established yet. This includes functionality related to the recognition of different operational regimes, as well as their use in the inferential measurement of froth grade. The last decade has seen major breakthroughs in deep learning and advances in image processing, which have also had a direct impact on flotation froth image analysis with computer vision systems. In this paper, these advances are reviewed and future trends are identified. Convolutional neural networks that are able to learn features from froth images have redefined the state-of-the-art in froth image analysis. These models rely heavily on transfer learning, with models such as GoogLeNet and MobileNet leading in the field. Emerging trends comprise a stronger focus on dynamic froth image analysis or the analysis of froth video sequences, froth-based monitoring, exploitation of froth features in advanced control and one-shot learning approaches based on froth image synthesis. Challenges are related to the labelling of images, the computational cost associated with training deep neural networks, as well as interpretation of these models. [ABSTRACT FROM AUTHOR]

Huge economic losses occur in the agricultural industry due to bacterial, viral or fungal infections in crops due to which farmers incur 15–20% losses on their profits annually. India is the second largest producer of rice and a leading exporter of the same in the global market. Thus, early disease detection in crops is essential. Implementing Smart Farming is a burning area of research in order to prevent further damage to crops. The widespread development of Deep Learning makes it possible to achieve the goal of disease detection in crops. In this paper we have proposed an intelligent model based on Smart farming integrating Machine Learning with the IoT network. The novelty of this project is early detection of Brown spot disease in rice paddy for the very first time using Convolutional Neural Networks. Deep Learning uses Neural Networks to implement Artificial Intelligence. This project makes use of Image recognition and pre-processing based on real time data. Data pre-processing and feature extraction has been done using a self-designed image-processing tool. Tensor flow and Keras framework has been implemented on both training and testing data collected manually from rice fields. The proposed model achieved an accuracy of 97.701% posing the ability to minimise the losses overall to the national and global productions. Further an app has been designed for the farmers for the same. [ABSTRACT FROM AUTHOR]

This article presents a state-of-the-art review of the applications of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) in building and construction industry 4.0 in the facets of architectural design and visualization; material design and optimization; structural design and analysis; offsite manufacturing and automation; construction management, progress monitoring, and safety; smart operation, building management and health monitoring; and durability, life cycle analysis, and circular economy. This paper presents a unique perspective on applications of AI/DL/ML in these domains for the complete building lifecycle, from conceptual stage, design stage, construction stage, operational and maintenance stage until the end of life. Furthermore, data collection strategies using smart vision and sensors, data cleaning methods (post-processing), data storage for developing these models are discussed, and the challenges in model development and strategies to overcome these challenges are elaborated. Future trends in these domains and possible research avenues are also presented. [Display omitted] • AI, ML and DL are transforming the building and construction industry 4.0. • Robotics, IoT, Digital Twins, 4D Printing, Cloud AR/VR, Blockchain are future trends. • AI techniques can be used throughout the whole building lifecycle. • Sensor and vision-based data acquisition can facilitate ML/DL implementation. [ABSTRACT FROM AUTHOR]

Deep learning and neural networks are being extended to geometric design and computation. Recent studies show that deep neural networks are vulnerable to adversarial samples. However, for 3D mesh adversarial samples, the most related studies actually attack the 2D victim networks, in which they have to project 3D objects to 2D images. In this paper, we present D3AdvM (Direct 3D adversary for mesh) to directly generate adversarial samples inside mesh data. Specifically, we propose two adversary generation approaches: vertex-based and edge-based. The first one, 3DVP (3D Vertex-based Perturbation), skillfully searches the optimized vertices positions based on opposite gradient fitness. The second one, KES (Key Edge-based Selection), carefully collapses the key feature edges according to importance of edge feature. Thus, our approaches avoid 3D/2D projections and approximation errors. Also, our approach can reduce the computation overhead. In addition, the proposed D3AdvM can control the number of perturbed vertices for real-world engineering designs and applications. Extensive experiments show that the generated 3D meshes are effective to attack classification networks. Furthermore, we evaluate the transferability, in which D3AdvM can attack both mesh-based networks and point cloud-based networks as victim networks. Our findings could benefit 3D geometry design based on the new generation of artificial intelligence and big data. • This work is the first one to directly attack mesh-based neural networks with adversarial meshes in 3D space without the 2D projection. • We explore the characteristics of 3D mesh topological entities and propose both vertex-based and edge-based adversarial approaches. • D3AdvM can constrain the adversarial samples within a specific number to support high feasibility in real-world applications. • D3AdvM shows valuable transferability to attack different versions of mesh-based networks and even the point cloud-based networks. • Our findings could benefit 3D geometry design based on the new generation of artificial intelligence and big data. [ABSTRACT FROM AUTHOR]

• The necessity of radiographic image segmentation is described. • The commonly used deep learning algorithms for image segmentation are reviewed. • The feasibility of image segmentation using deep learning network model is discussed. • The advantages and disadvantages of neural network models are compared. • The review found that deep learning models have good performance in image segmentation. Due to the advancement of medical imaging and computer technology, machine intelligence to analyze clinical image data increases the probability of disease prevention and successful treatment. When diagnosing and detecting heart disease, medical imaging can provide high-resolution scans of every organ or tissue in the heart. The diagnostic results obtained by the imaging method are less susceptible to human interference. They can process numerous patient information, assist doctors in early detection of heart disease, intervene and treat patients, and improve the understanding of heart disease symptoms and clinical diagnosis of great significance. In a computer-aided diagnosis system, accurate segmentation of cardiac scan images is the basis and premise of subsequent thoracic function analysis and 3D image reconstruction. This paper systematically reviews automatic methods and some difficulties for cardiac segmentation in radiographic images. Combined with recent advanced deep learning techniques, the feasibility of using deep learning network models for image segmentation is discussed, and the commonly used deep learning frameworks are compared. There are many standard methods for medical image segmentation, such as traditional methods based on regions and edges and methods based on deep learning. Because of characteristics of non-uniform grayscale, individual differences, artifacts and noise of medical images, the above image segmentation methods have certain limitations. It is tough to obtain the needed results sensitivity and accuracy when performing heart segmentation. The deep learning model proposed has achieved good results in image segmentation. Accurate segmentation improves the accuracy of disease diagnosis and reduces subsequent irrelevant computations. There are two requirements for accurate segmentation of radiological images. One is to use image segmentation to improve the development of computer-aided diagnosis. The other is to achieve complete segmentation of the heart. When there are lesions or deformities in the heart, there will be some abnormalities in the radiographic images, and the segmentation algorithm needs to segment the heart altogether. The quantity of processing inside a certain range will no longer be a restriction for real-time detection with the advancement of deep learning and the enhancement of hardware device performance. [ABSTRACT FROM AUTHOR]

In this paper, we propose a deep neural network model to simulate the transient ultrasonic wave propagation in the 2D domain by implementing the Data driven-simulation-assisted-Physics learned AI (DPAI) model. The DPAI model consists of modified convolutional long short-term memory (ConvLSTM) with an encoder–decoder structure, which learns the representation of spatio-temporal dependence from input sequence data. The DPAI uses the data-driven approach to understand the underlying physics of elastic wave propagation in a medium. This model is trained with simulation-assisted finite element simulation datasets consisting of distributed single and multi-point excitation sources in the medium. The effectiveness of the proposed approach is demonstrated by modeling a wide range of scenarios in elastodynamic physics, such as multiple point sources, varying excitation parameters, and wave propagation in a large 2D domain. The trained DPAI model is tested and compared against FE modeling with respect to accuracy and computational time. • DPAI model applicable for solving conventional ultrasonic wave propagation problems in the 2D domain. • Trained DPAI can generate wave propagation simulations for different frequencies, the number of cycles, excitation sources. • DPAI is efficient in computational time and outperforms FE modeling with adequate computational resources. • The same DPAI approach needs to be extended to simulate the reflection and mode conversion phenomenon. [ABSTRACT FROM AUTHOR]

In the last few years, we have witnessed a renewed and fast-growing interest in continual learning with deep neural networks with the shared objective of making current AI systems more adaptive, efficient and autonomous. However, despite the significant and undoubted progress of the field in addressing the issue of catastrophic forgetting, benchmarking different continual learning approaches is a difficult task by itself. In fact, given the proliferation of different settings, training and evaluation protocols, metrics and nomenclature, it is often tricky to properly characterize a continual learning algorithm, relate it to other solutions and gauge its real-world applicability. The first Continual Learning in Computer Vision challenge held at CVPR in 2020 has been one of the first opportunities to evaluate different continual learning algorithms on a common hardware with a large set of shared evaluation metrics and 3 different settings based on the realistic CORe50 video benchmark. In this paper, we report the main results of the competition, which counted more than 79 teams registered and 11 finalists. We also summarize the winning approaches, current challenges and future research directions. [ABSTRACT FROM AUTHOR]

In the last decades, research has claimed facial micro-expressions are a reliable means to detect deception. However, experimental results showed that trained and naïve human judges observing facial micro-expressions can distinguish liars from truth-tellers with an accuracy just slightly above the chance level. More promising results recently came from the field of artificial intelligence, in which machine learning (ML) techniques are used to identify micro-expressions and are trained to distinguish deceptive statements from genuine ones. In this paper, we test the ability of different feature extraction methods (i.e., improved dense trajectories, OpenFace) and ML techniques (i.e., support vector machines vs. deep neural networks) to distinguish liars from truth-tellers based on facial micro-expressions, using a new video data set collected in low-stakes situations. During the interviews, a technique to increase liars' cognitive load was applied, facilitating cues of lies to emerge. Results highlighted that support vector machines (SVMs) coupled with OpenFace resulted in the best performing method (AUC = 0.72 videos without cognitive load; AUC = 0.78 videos with cognitive load). All the tested classifiers performed better when a cognitive load was imposed on the interviewee, confirming that the technique of increasing cognitive load during an interview facilitates deception recognition. In the same task, human judges obtained an accuracy of 57%. Results are discussed and compared with the previous literature, confirming that artificial intelligence performs better than humans in lie-detection tasks do, even when humans have more information to make a decision. • Automatic systems are better than humans in detecting deception from videos. • AI is a promising methodology to detect deception through facial micro-expressions. • AI performs better when the cognitive load of the interviewees is increased. • SVM is more able in distinguish liars from truth-tellers when trained with AUs as features. [ABSTRACT FROM AUTHOR]

In this paper, the twin-systems approach is reviewed, implemented, and competitively tested as a post-hoc explanation-by-example solution to the eXplainable Artificial Intelligence (XAI) problem. In twin-systems, an opaque artificial neural network (ANN) is explained by "twinning" it with a more interpretable case-based reasoning (CBR) system, by mapping the feature weights from the former to the latter. Extensive comparative tests are performed, over four experiments, to determine the optimal feature-weighting method for such twin-systems. Twin-systems for traditional multilayer perceptron (MLP) networks (MLP–CBR twins), convolutional neural networks (CNNs; CNN–CBR twins), and transformers for NLP (BERT–CBR twins) are examined. In addition, Feature Activation Maps (FAMs) are explored to enhance explainability by providing an additional layer of explanatory insight. The wider implications of this research on XAI is discussed, and a code library is provided to ease replicability. [ABSTRACT FROM AUTHOR]

• A new deep learning architecture that integrates spiking neural network (SNN) for contextual information extraction is proposed. • The proposed design shows improved robustness for both random and structured input perturbation during inference. • SNN is implemented with a novel frequency-dependent stochastic spike-timing-dependent-plasticity learning rule. Spiking neural network (SNN) is a type of artificial neural network that uses biologically inspired neuron models and learning rules to develop artificial intelligence with capability parallel to human brain. Deep neural networks (DNNs), on the other hand, uses less biologically plausible neurons and training methods such as gradient descent, and has shown good accuracy in computer vision tasks. However, human brain can still outperform DNN in certain scenarios. For example, DNN experiences significant performance degradation when perturbation from various sources is present in the input, which makes DNN less reliable for systems interacting with physical world. In this paper, we present a hybrid deep net work architecture with s pike-assisted c ontextual i nformation e xtraction (ScieNet) as a solution to the problem. ScieNet integrates a front-end SNN with a novel stochastic spike-timing-dependent plasticity (STDP) algorithm that extracts visual context from images. The back-end DNN is trained for classification given the contextual information. The integrated network demonstrates high resilience to input perturbations without relying on pre-training on perturbed inputs. We demonstrate ScieNet with various back-end DNNs for image classification using different datasets and considering both stochastic and structured input perturbations. Experimental results demonstrate significant improvement in accuracy on perturbed images, while maintaining state-of-the-art accuracy on clean images. [ABSTRACT FROM AUTHOR]

Visual design is associated with the use of some basic design elements and principles. Those are applied by the designers in the various disciplines for aesthetic purposes, relying on an intuitive and subjective process. Thus, numerical analysis of design visuals and disclosure of the aesthetic value embedded in them are considered as hard. However, it has become possible with emerging artificial intelligence technologies. This research aims at a neural network model, which recognizes and classifies the design principles over different domains. The domains include artwork produced since the late 20th century; professional photos; and facade pictures of contemporary buildings. The data collection and curation processes, including the production of computationally-based synthetic dataset, is genuine. The proposed model learns from the knowledge of myriads of original designs, by capturing the underlying shared patterns. It is expected to consolidate design processes by providing an aesthetic evaluation of the visual compositions with objectivity. [Display omitted] • This paper examines quantitative aspects of visual design principles (71). • A CNN based AI model is developed to detect fundamental visual features (74). • 9 classes -stem from 3 main principles: emphasis, balance and rhythm are targeted (84). • Visual data from photography, painting and architecture domains are used for ML (82). • High accuracy rates are recorded for top 3 predictions among 5 experiment settings (85). [ABSTRACT FROM AUTHOR]

Recent breakthroughs in artificial intelligence (AI), and particularly in deep learning (DL), have created tremendous excitement and opportunities in the earth and environmental sciences communities. To leverage these new 'data-driven' technologies, however, one needs to understand the fundamental concepts that give rise to DL and how they differ from 'process-based', mechanistic modelling. This paper revisits those fundamentals and addresses 10 questions that might be posed by earth and environmental scientists, and with the aid of a real-world modelling experiment, it explains some critical, but often ignored, issues DL may face in practice. The overarching objective is to contribute to a future of AI-assisted earth and environmental sciences where AI models can (1) embrace the typically ignored knowledge base available, (2) function credibly in 'true' out-of-sample prediction, and (3) handle non-stationarity in earth and environmental systems. Comparing and contrasting earth and environmental problems with prominent AI applications, such as playing chess and trading in stock markets, provides critical insights for better directing future research in this field. • DL is rooted in connectionism, hyper-flexibility, and vigorous optimization , which are alien to process-based modelling. • A knowledge base is essential to enable predictions of complex, open , partially observable , and non-stationary systems. • Bridging DL and process-based modelling is embryonic but has great potential in an age of big data and computational power. [ABSTRACT FROM AUTHOR]