Abstract

Deep learning techniques have shown significant impact in enhancing security across various domains by leveraging artificial neural networks models. When applied to cloud computing security, deep learning offers cost-effective solutions by automating threat detection, reducing manual monitoring, and improving overall security effectiveness. Deep learning models using neural networks play crucial role in tasks like intrusion detection, malware detection, anomaly detection, and log analysis. Integration of deep learning into cloud security requires careful evaluation of existing systems, defining objectives, dataset selection and preparation, model tuning, and eventual modifications for compatibility. Furthermore, implementing deep learning techniques in cloud security entails considering factors such as computational resources, data collection and preparation costs, model development, integration efforts, and ongoing monitoring and maintenance. This paper proposes a feed-forward propagation Artificial Neural Network (ANN) model in cloud security and investigates the key steps for integrating such models into cloud security strategies. Considering that the effectiveness of the ANN model depends on factors such as training data quality, network architecture, and weight adjustment algorithms, the study utilizes a dataset from Kaggle.com for validation and demonstrates steps involved in training and evaluation of the ANN model.

Introduction

Deep learning (DL) techniques have emerged to become effective tools for enhancing security in various domains. These techniques leverage the capabilities of artificial neural networks to learn and identify patterns from vast amounts of data, enabling more robust and efficient security solutions [1][2]. When applied to cloud computing security, deep learning techniques offer several benefits having direct implications in the cost-effectiveness of the products. A great benefit of implementation of DL in security of cloud is seen in automated threat detection. DL algorithms can analyze large volumes of data, such as network traffic logs, system logs, and user behavior, to detect anomalies and potential security threats automatically [2], [3]. This together with other utilizations as a result of DL, reduces the need for manual monitoring and analysis, what leads to faster identification of security incidents, timely response, mitigation, and reduced costs in time and resources. Considering this, through automation of various security tasks, DL minimizes the risk of human error, which can lead to security breaches and associated costs.

Automated systems powered by deep learning can perform tasks consistently and accurately, enhancing overall security effectiveness [4], [5]. DL models learn patterns and relationships in data, allowing for more accurate threat detection and classification. Through detecting patterns and anomalies which could go unnoticed by traditional rulebased systems, DL implementation makes it possible for organizations to optimize resource allocation and minimize unnecessary costs.

Conclusion

The integration of deep learning techniques into cloud computing security offers numerous benefits. By analyzing large volumes of data, deep learning algorithms can detect security threats in real-time and minimize various risks. However, the implementation of deep learning in cloud security requires careful evaluation of costs and resources. Organizations struggle when having to adjust the solution to the computational resources, data acquisition and preparation costs, personnel expertise, and ongoing monitoring and maintenance.

The integration of deep learning techniques in cloud security involves few steps for a successful outcome. These include assessing the strengths and weaknesses of the existing system, identifying specific tasks that deep learning can address, selecting suitable models and algorithms, training the models with labeled data, and ensuring compatibility and integration with existing components. Continuous monitoring and feedback are also crucial stages for maintaining the efficiency and effectiveness of the integrated system.

Abstract

This study presents a novel financial performance forecasting method that combines the threshold technique with Artificial Neural Networks (ANN). It applies the threshold regression method to identify the factors within the board of directors that influence the financial performance of traditional industries in Taiwan. The findings indicate that the ANN method effectively predicts financial performance by using relevant board structure data. Furthermore, the empirical results suggest that boards with more members demonstrate increased profitability. Additionally, a more significant presence of board members with accounting expertise contributes to more consistent profits. In contrast, an increased presence of members with financial expertise has a more pronounced impact on profitability.

Introduction

Traditional industries, including sectors such as metal machinery, the chemical industry, and subsistence industries, have played pivotal roles in Taiwan’s economic development, trade, and industrial evolution over the years. However, in recent years we have witnessed the emergence of economic powerhouses and the rapid progress of regional economic and trade partnerships like ASEAN, resulting in significant changes to the global economic and trade landscape. As a result, these industries are now operating within a new and highly competitive environment. In this era of rapid transformation, these industries must identify the essential factors that will enable them to maintain and enhance their positions on the international stage.

Investors seek to achieve multiple objectives through the election of a company’s board of directors. They aim to enhance corporate governance, safeguard the rights and interests of shareholders, and anticipate that the board will contribute expertise to improve the decision-making prowess of the board itself. Ultimately, this is expected to enhance the company’s overall performance. Consequently, whether the board of directors can positively impact corporate performance has remained a significant research focus in management and academia.

While many studies have explored the relationship between the board of directors’ characteristics and company performance, a notable need exists for more in-depth articles examining the professional backgrounds of the board members and CEOs and how their leadership qualities influence financial performance. This gap in research underscores the need for a more comprehensive understanding of the intricate dynamics between leadership, professional expertise, and corporate financial outcomes.

Conclusions

The research results show that, for companies in traditional industries that are listed on the Taiwanese stock market, the composition of the board of directors impacts the business’s financial performance.

Boards with more members may have a greater capacity for overseeing corporate governance. A greater representation of board members with accounting expertise tends to lead to more consistent corporate profits, whereas an increased presence of members with financial expertise results in more pronounced profit fluctuations. Additionally, in the case of companies with lower profits, managers may face pressure from loan contracts or receive limited attention from investors, potentially leading to heightened practices related to profit management. Managers may aim to maximize their benefits. At this time, the return on assets is high, and the managers obtain the expected benefits. Thus, to restrain the profit management behavior of managers, enterprises need to have a broader view and have specific judgments and analyses in each enterprise to build organizational structure and procedures. The operation of the board of directors is more effective because of the industry’s characteristics, the enterprise’s size, the capital structure, and the performance of the enterprise. Finally, this study used deep learning and big data techniques to build an Artificial Neural Network (ANN) model that specializes in predicting the ROA of companies based on most factors related to board structure, with a prediction accuracy of over 78%.

Abstract

     At present, intelligent computing applications are widely used in different domains, including retail stores. The analysis of customer behaviour has become crucial for the benefit of both customers and retailers. In this regard, the concept of remote gaze estimation using deep learning has shown promising results in analyzing customer behaviour in retail due to its scalability, robustness, low cost, and uninterrupted nature. This study presents a three-stage, three-attention-based deep convolutional neural network for remote gaze estimation in retail using image data. In the first stage, we design a mechanism to estimate the 3D gaze of the subject using image data and monocular depth estimation. The second stage presents a novel three-attention mechanism to estimate the gaze in the wild from field-of-view, depth range, and object channel attentions. The third stage generates the gaze saliency heatmap from the output attention map of the second stage. We train and evaluate the proposed model using benchmark GOO-Real dataset and compare results with baseline models. Further, we adapt our model to real-retail environments by introducing a novel Retail Gaze dataset. Extensive experiments demonstrate that our approach significantly improves remote gaze target estimation performance on GOO-Real and Retail Gaze datasets.

Introduction

     In today’s world, retail stores are becoming smarter with the availability of numerous data and the power to analyze them autonomously. Even with the rise of online shopping, most of the physical retail stores use smart applications for the purchasing process [1]. Several techniques and devices have been introduced to automate the shopping process and analyze shoppers’ behaviour inside stores. At the same time, the shopping experience is a key consideration towards the success of a retail business, which affects the performance of customer satisfaction, customer purchase probability, and customer loyalty [2]–[3][4].

     In order to improve the shopping experience and maximize business profits, it is essential to capture and and analyze the customer’s behaviours without interfering their natural shopping journey [5], [6]. Various solutions have introduced for customer behaviour analyzis in retail using developments in computer vision technology. For instance, counting the number of people and detecting the hot spots in retail [6] and public [7], and tracking shoppers’ emotion [5] are such applications. However, the existing solutions only capture coarse touch-points of a shopper’s journey and vulnerable to unconstrained environment settings. With the adaptation of computer vision technologies in gaze estimation, there has been eye tracking-based solutions for customer behaviour analysis in retail as well [1], [8]. Moreover, there are solutions based on virtual reality devices and head-mounted displays, wearable eye tracker based solutions [9], and non-intrusive 3D eye tracking solutions [10]. However, these solutions do not completely satisfy the retailers due to high cost of 3D eye tracking solutions, unscalability of wearable, and head-mounted display-based solutions, and manual calibration of eye tracking systems.

Conclusion

     Remote gaze saliency estimation in retail is a novel concept that has a significant potential towards building innovative retail stores. In this study, we researched the application of remote gaze saliency estimation for non-interruptive, low-cost, and scalable customer behaviour analysis in retail. We proposed a Depth-based Dual Attention model, a three-stage, three-attention-based deep CNN for gaze saliency estimation from back-head images in the wild. We developed four design solutions to comprehensively represent the parameters of gaze saliency estimation problem in retail and introduced the novel object channel and depth-rebasing components as hand-designed features, designed in our two preceding model architectures and then combined in the final model.

     Extensive quantitative and qualitative analysis on the benchmark GOO-Real dataset demonstrates the superiority of the proposed models and the importance of our introduced hand-designed components. Our proposed solution improved 33% for angular error compared to the current best work in the literature. Furthermore, we introduced Retail Gaze, a real-world retail gaze saliency estimation dataset, to ensure the validity and applicability of our proposed solution in real retail environments. The proposed solution achieved an angular error of 15.3° on the Retail Gaze dataset, which demonstrates that it performs favourably in real retail environments.

Abstract

     This work proposes a smart crop growth monitoring system that contains an adaptive cryptography engine to ensure the security of sensor data and an edge artificial intelligence (AI) based estimator to classify the pest and disease severity (PDS) of target crops. Based on the smart system management mechanism, cryptographic functions can be adapted to varying and real-time requirements, while the actuators can be controlled to interact with the physical world to ensure the healthy growth of crops. Experiments show when all the four cryptographic hardware modules, including RTEA32, RTEA64, XTEA32 and XTEA64, are supported, using the adaptive cryptography engine, 72.4% of slice LUTs and 68.4% of slice registers in terms of the Xilinx Zynq-7000 XC7Z020 chip can be saved. Through the smart system management mechanism, a power consumption of 0.009 watts can be reduced. Furthermore, using the binarized neural network (BNN) hardware module of the PDS estimator, the recognition accuracy of target crops i.e. dragon fruits can achieve 76.57%. Compared to the microprocessor-based design and the GPU accelerated one, the same BNN architecture on the FPGA can accelerate the frames per second by a factor of 4,919.29 and a factor of 1.08, respectively.

Introduction

     Recently, the abnormal climate leads to the extreme weather, while the occurrence of natural disasters such as typhoon, rainstorm and severe drought gradually increases. This causes great casualties and serious damages to our properties and environment. For agriculture, the extreme weather also makes the growth of crops unstable, and the problem of food shortage thus becomes more and more serious. For all countries in the world, the food crisis has also become a very important issue.

     Until now, most crops are still planted in the outdoor. This means the growth of crops will be affected by the weather easily. This also makes the yield and quality of farm crops unstable. Compared to the opening planting environments, recently, the greenhouse becomes a new alternative due to its controllable advantage. With the incoming of agriculture 4.0, new techniques such as cyber physical systems (CPS) [1] and Internet-of-Things (IoT) [2] further enhance the efficiency of the agricultural management. Furthermore, with the popularity of big data analytics [3], the trend of crop growth can be predicted and analyzed. For example, by applying sensors to the planting environment of crops, the collected data can be further analyzed to improve the productivity and quality of crops. Furthermore, the corresponding actuators such as sprinklers can be also controlled to interact with the physical world to ensure the healthy growth of crops.

Conclusion

     To monitor the crop growth efficiently, this work proposes a crop growth monitoring system based on system adaptivity and edge AI. The presented adaptive cryptography engine can not only support varying requirements of cryptographic functions but also provide real-time decryption processing of sensor data. Furthermore, the layered and virtualizable design makes the crop growth monitoring system scalable. The edge AI based PDS estimator provides real-time detection of the target crops, while the image fusion method can assist in classifying the level of PDS. Through the smart system management mechanism along with the adaptive cryptography engine and the PDS estimator, the actuators can be controlled to interact with the physical world to ensure the healthy growth of crops. Our experiments also demonstrated the practicability and applicability of the proposed design.

Abstract

     This paper presents an Artificial Nerual Network (ANN) for identification of postmenopausal women who are at high risk for developing osteopathy. While 800 patients took part in the study, 180 were used for network training. The following parameters were used: T-score (from −2,5 to −4), Age, Blood calcium level (<1,9 mmol/L), Blood vitamin D level (<20 ng/ml), Hip fracture, Spine fracture, Joint fracture, Glucocorticoids use, Smoking status, and BMI. The network has 10 input parameters and 1 output parameter. For the final architecture of expert system, a neural network with 20 neurons in hidden layer was chosen based on the training results. The signal from each neuron from hidden layer is directed to neuron in output layer, where this neuron processes the signal and gives desired output of the network. The sensitivity was 97,5%, specificity 70%, and accuracy 94,44%.

Introduction

     Osteoporosis is a skeletal illness that causes weakening of the bones, which can lead to increased fractures. People with osteoporosis have decreased bone mass and microarchitectural degeneration of bone tissue, in addition to lower bone strength [1-3].

     Osteoporosis is a skeletal illness that causes weakening of the bones, which can lead to increased fractures. People with osteoporosis have decreased bone mass and microarchitectural degeneration of bone tissue, in addition to lower bone strength [1-3].

     According to data from 2010, 6,6% of men and 22,1% of women over 50 in the EU have osteoporosis. Because of the expanding number of patients worldwide, it is now referred to as a “silent epidemic.” Variable and non-variable factors that raise the chance of developing osteoporosis and bone fractures can be separated. One of the most important variables is smoking, which is linked to decreased bone resistance to mechanical stresses and friction. The effects of excessive alcohol use on bone homeostasis are significant. Caffeine, glucocorticoid therapy, insufficient calcium and vitamin D intake, insufficient physical activity, low BMI, past bone fractures, and a family history of osteoporosis are all risk factors [4-6].

Conclusion

     As osteoporosis in postmenopausal women is a rising problem in the modern world, predictive modelling of the risk of developing osteoprosis is highly desirable. This paper presents the development and validation of an ANN based model for prediction and automated diagnosis of osteoporosis in post-menopausal women based on risk-contributing factors. Giving the high accuracy and sensitivity of proposed ANN for identification of high risk of osteoporosis development in postmenopausal women, it can be concluded that AI has a high potential for decission making for this specific purpose. Prediction of high risk for osteoprosis development can contribute to adjustments in lifestyle and possible prevention of osteoporosis. In addition to the benefit this would have to each individual, the cost reduction in terms of preventing costly interventions necessary in case of osteoporosis development is a significant contribution.

Abstract

References

Acknowledgements

Author information

Ethics declarations

Additional information

Rights and permissions

About this article

Abstract

The purpose is to make full use of data mining and machine learning technology under big data to improve the ability of trade financial enterprises to cope with the risk of excessive financialization. In view of the above needs, based on previous studies, genetic algorithm (GA), neural network and principal component analysis (PCA) methods are used to collect and process the data, and build a risk assessment model of excessive financialization of financial enterprises. The performance of the model is analyzed through the data of specific cases. The results suggest that the data mining technology based on back propagation neural network (BPNN) can optimize the input variables and effectively extract the hidden information from the data. The specific examples show that most of the current enterprises do not have greater financial risk. However, most of the financial enterprise indexes show that the actual enterprise assets are gradually financialized. The total accuracy rate of financial risk assessment model based on deep belief network (DBN) is over 91%, and the accuracy of the model can reach 80% even if the sample size is small. Therefore, the financial risk assessment model proposed can effectively analyze the relevant financial data, and provide reference for the financial decision-making research of financial enterprises.

1 Introduction

With the continuous development of the world economy in recent years, more and more enterprises are active in the fnancial market, which promotes the sustainable development of the national economy. However, the following problems are the fracture of many enterprise capital chains due to the poor risk awareness, fnancial market turbulence, commercial fraud or mismanagement. Small enterprises are on the verge of bankruptcy because of their poor ability to resist risks, which is mainly because more enterprises do not invest in the real economy but pay more attention to virtual fnancial markets (Anginer et al., 2018). With the advent of big data era, although the risk control model of enterprises plays a certain role in credit management, due to the lack of deep understanding of data, the corresponding model accuracy is low, and it is unable to evaluate the borrower comprehensively (Florio & Leoni, 2017). Due to the more in-depth research on big data tools by many scholars, more and more methods are applied to the fnancial feld (Trelewicz, 2017). Traditional risk control model has problems such as single dimension and limited assessment ability. However, using data mining method can analyze the data of fnancial industry in depth, involving many dimensions and more comprehensive assessment (Moradi & Mokhatab Rafei, 2019). With the establishment of distributed database and the gradual maturity of various data platform architectures, more data are stored and calculated, and the dimensions of data analysis are also wider (Medvedev et  al., 2017). Especially with the construction of Hadoop cloud computing platform, the distributed computing ability of big data has been greatly improved, and the data foundation and computing ability are more efcient. Compared with the traditional risk control model, the model based on big data has great changes in the amount of information and accuracy (Ding et al., 2019; Ivanov et al., 2019). Big data can improve the credit system and help trade enterprises reduce credit risk (Saura et  al., 2019; Urbinati et  al., 2019). Data mining can use the network data to improve the scoring model and optimize the approval process through the integration of algorithms, so as to form a sustainable closed-loop management mode (Aldridge, 2019). Therefore, the use of data mining and its application in fnancial risk assessment has become one of the hot spots in this feld.

1. Introduction and research motivation

2. Research on Doom using the VizDoom Game-AI research platform

3. Direct future prediction (DFP)

4. Direct future prediction (DFP) and asynchronous advantage Actor-Critic (A3C)

5. Advantage Actor-critic (A2C) and advantage Actor-Critic- long Short-Term memory (A2C-LSTM)

6. Asynchronous advantage Actor-Critic (A3C) and deep recurrent Q-Network (DRQN)

7. Direct future prediction (DFP) and deep recurrent Q-Network (DRQN)

8. Deep Q-Network (DQN) and deep recurrent Q-Network (DRQN)

9. Dueling deep Q-Network (DDQN) and double deep Q-Network (DDQN)

10. C51_DDQN and Reinforce

11. Techniques with better performance

12. Discussion and future work

Declaration of Competing Interest

Acknowledgment

References

Abstract

Artificial Intelligence in the form of machine learning is employed in games to control non-human computer-players, agents or bots. However, most of these games such as Atari took place in 2D environments that were fully observable to the agents. Currently, it is of extreme significance to employ such machine learning techniques and methods in 3D environments such as Doom. Therefore, In this paper, we train agents on the health gathering scenario of the classical first-person shooter game Doom by first presenting the Direct Future Prediction to train an agent that uses a simple architecture with no additional supervisory signals, then differentiate and compare the performance of the agents trained by using several different machine learning techniques, and the AI reinforcement learning platform ‘VizDoom’, a 3D partially observable environment, with interesting enhanced properties that makes agents to stand out from inbuilt AI agents and human players. We have continued to use computer games as a benchmark for the performance of AI as having been so successful in the past. We also compared the results of our findings to conclude the performance of the agents trained with different machine learning techniques. The agents performed well against both human players and inbuilt game agents.

Introduction and research motivation

In the last few decades, due to the progress in artificial intelligence, a revolution and sudden change have been observed in the technology both in hardware and software [1]. This change is seeping and taking over in our lives up to a certain extent, affecting how we live, work and entertain ourselves such as employing domestic robots servants, healthcare uses, electronic trading, remote sensing, expert systems, traffic control systems, autonomously-powered self-driving vehicles, and from behavioral algorithms to suggestive searches [2], etc. In the same way, gaming is a widely recognized part of our cultural landscape and as old as our human ancestors. The earliest computers were very slow and the interaction with the user was limited to basic principles. In the early ’40s, computers evolved, and programmers started to develop new virtual worlds and surprising ways of interaction between the user and the machine [3]. But now due to advancements in technology such as GPU’s[4], TPU’s[5] and the revolution in deep neural networks [6] it has become possible for artificial intelligence to stepin in video games as well where massive graphical data in the form of frames, or to be more specific a huge amount of multidimensional data is required to be processed and execute [7]. In the recent past, machine learning techniques and methods were employed in Atari games for training agents, where later, the agents performed on 49 different Atari games with better and improved results. However, most of these Atari games took place in 2D environments that were fully observable to the agents [8]. Currently, it is of extreme significance to employ such machine learning techniques and methods in 3D environments such as Doom [9] a first-person-shooter game shown in fig. 1, Starcraft [10] a third person shooter game based on real-time strategies, and sandbox open-world games such as Grand Theft Auto V and Minecraft [11] because the research community in AI think and consider that computer video games are the best test-beds for testing different artificial intelligence techniques, methods, and algorithms before evaluating them in real-world life. Thus, in this paper, state-of-the-art machine learning techniques that were before partially tested in 2D environments are now employed in a 3D environment known as Doom, to train, differentiate and compare agents performances, such as advantage actor-critic (A2C) [12], advantage actor-critic long short-term memory (A2C-LSTM) [13], asynchronous advantage actor-critic (A3C) [14], Deep Q-network (DQN) [15], Deep recurrent Q-network (DRQN) [16], Double deep Q-network (DDQN) [17], C51-DDQN [18], Dueling deep Q-network (DDQN) [19], and Reinforce [20] whereafter applying most of the agents are found useful and effective. In addition, this paper presents one of the 4 best techniques that performed well on the VizDoom AI platform [21]. It was suggested that making such research available is beneficial for the community researching on first-person-shooter games which may set up a base for further research and improvement.

1- Introduction

2- Methods and materials

3- Results

4- Discussion

5- Conclusion

References

Abstract

Atrial fibrillation (AF) is a common cardiac arrhythmia in clinic. The traditional AF detection using visual inspection by trained physicians is an inefficient and burdensome task. In this paper, we introduce a novel method for the automated AF detection using two-lead electrocardiogram (ECG) signals. We use the wavelet packet transform (WPT) and the correlation function of random process theory to devise an efficient feature extraction strategy for physiological signal analysis, and construct the corresponding histogram. Then, multivariate statistical features based on the correlation among wavelet coefficient series are extracted from the corresponding histogram as the feature set, which is the input to artificial neural network (ANN) classifier for the detection. Moreover, various statistical analyses are performed and some parameter tuning strategies are formulated by fitting the receiver operating characteristic (ROC) curve to ensure the reliability and robustness of the work. To evaluate the classification performance of the algorithm, 10-fold cross-validation is implemented on the MIT-BIH AF database. Compared with some state-of-the-art algorithms, the numerical results prove that our proposed strategy yields superior classification performance. To the best of our knowledge, this is also the first application of random process theory for AF detection, providing great potential in medical diagnosis.

Introduction

Atrial fibrillation (AF), a common cardiac arrhythmia, endangers human health and causes a social burden, which is closely related to the occurrence of stroke, thrombus, even death. More than 6 million people have been affected in the United States and about 90 million people have AF around the world [1]. This trend is expected to increase significantly with age [2]. Therefore, it is essential for public health to combat the harm and prevalence of the disease. Conventionally, AF detection is diagnosed by visual inspection of electrocardiogram (ECG) signals by trained physicians, whichmakes artificialdetectioninefficient andsubjective [3]. A large amount of ECG data have inevitably hindered the efficiency of AF detection. Hence,there is an urgent demand for an automated AF detection mechanism to analyze massive amounts of ECG data, facilitate diagnosis and lighten the burden on physicians [4]. The ECG is one of the most common and powerful tools for diagnosing atrial activity in clinicaltreatment, which mainly reveals the electrical action in the heart of human body [3,5]. The irregularity of RR intervals and absence of P-waves (replaced by rapid, irregular and disordered fibrillatory waves, called f-waves) are two main features of ECG data in AF [6]. Various algorithms based on these two features have been performed to automatically detect AF from ECG data. According to the absence of P-waves, Maji et al. [7] utilized empirical mode decomposition to analyze the denoised ECG signals for the corresponding intrinsic mode functions, and then some statistical parameters of P-waves were derived from these functions to perform classification. Padmavathi and Ramakrishna [8] measured some related coefficients based on the auto-regressive model of ECG signals and the support vector machine (SVM) as well as K-nearest neighbor (KNN) classifier was used to detect AF episodes. Pourbabaee et al.[9] developed a feature learning method using deep convolutional neural networks for the analysis of ECG data. At the same time, more schemes using the irregularity of RR intervals have also been initiated to classify AF signals. Dash et al.

1- Introduction

2- Materials and methods

3- Results and discussion

4- Conclusions

References

Abstract

Accelerated changes in land cover cause changes in environmental dynamics and may cause land degradation. The goals of the present paper were to analyze changes in land cover and to estimate a future scenario for 2035 using an artificial neural network for the Taperoá River basin, located in northeastern Brazil. The classification of land cover was carried out for years t1 (1990), t2 (1999) and t3 (2002), with the latter being used to validate the land cover prediction to obtain an estimate for year t4 (2035). The land cover classes identified in the basin were (a) water bodies, (b) tree-shrub vegetation, (c) shrub vegetation, (d) herbaceous-shrub vegetation, and (e) herbaceous vegetation. The results of the classifications and of the land cover prediction were analyzed using the kappa coefficient, total operating characteristic (TOC), and area under the curve (AUC). The dynamic modeling of the land cover was based on a multilayer perceptron (MLP) neural network, which presented very good results, with an accuracy = 89.69% after 10,000 iterations, kappa = 0.61 and AUC = 0.67. The results of the land cover change analysis showed a decrease in the tree-shrub class and an increase in the shrub vegetation class between the years analyzed. The scenario predicted for 2035 showed an increase in the herbaceous-shrub vegetation class and a decrease in the area occupied by tree-shrub vegetation.

Introduction

Changes in land cover generated by anthropogenic actions are responsible for changing landscape characteristics and modifying the dynamics of natural processes in river basins (Mendoza et al., 2011) and are of great interest to researchers due to their impact on the local and global environment (Abuelaish and Olmedo, 2016). The main causes of land cover changes vary according to the nature and extent of the area but include deforestation, changing to pasture, agricultural intensification, and overexploitation (Bezak et al., 2015). Spatial changes in land cover over time also affect the future provision and localization of ecosystem services in the landscape and the fragmentation of green areas (Hoyer and Chang, 2014; Bai et al., 2019). Changes in land cover pose threats to landscapes, for example, by removing characteristic features of the landscape by scale enlargement or deterioration due to lack of management (Schulp et al., 2019). The risk of these problems is even higher in the semiarid region of Brazil due to the lack of studies regarding the phenological behavior of the vegetation and the recurrence of droughts and rainfall variability, together with the geology of the region and the saline soil types (Silva et al., 2018a). The characteristics of the Caatinga biome make it difficult to produce forecasting maps of the vegetation cover due to the prolonged dry season in the region, which influences the spectral response of the vegetation and, consequently, the estimates of land cover changes (Santos et al., 2017).

1. Introduction

2. Quantum computing

3. Quantum computing approach for artificial neural networks (QC ANN)

4. Experimental evaluation

5. Conclusion and Future Scope

References

Artificial neural networks (ANN) proved to be efficient in solving many problems for big data analytics using machine learning. The complex and non-linear features of the input data can be learned and generalized by ANN. In the big data era, enormous amounts of data arrive from multiple sources. A stage is expected to be reached where even supercomputers are likely inundated with the big data. Training an ANN in such a situation is a challenging task due to the size and dimension of the big data. Also, a large number of parameters are to be used and optimized in the network to learn the patterns and analyze such data. Quantum computing is emerging as a field that provides a solution to this problem as a quantum computer can represent data differently using qubits. Qubits on quantum computers can be used to detect the hidden patterns in data that are difficult for a classical computer to find. Hence, there exists a huge scope for application in the area of artificial neural networks. In this work, we primarily focused on training an artificial neural network using qubits as artificial neurons. The simulation results show that our quantum computing approach for ANN (QC ANN) is efficient when compared to classical ANN. The model with qubits as artificial neurons can learn the features of data using fewer parameters for a binary classification task. We demonstrate our experiment using a quantum simulator and optimization of the quantum parameters used in QC ANN is carried out on a classical computer.

Introduction

Data processing is an area of concern, due to the rapid speeds at which the volume of data is growing. Getting some insights from very large amounts of data is a big challenge. IDC (International Data Corporation) in its white paper Data Age 2025 [15] predicts that, worldwide data will grow from 33ZB (1ZB = 1021bytes) in 2018 to 175ZB by 2025. Artificial neural networks [16, 12, 24] removed a lot of obstacles in big data processing, thereby, turning the data to information for successful decision making. Data growth and size create new problems to be addressed and challenge the existing computing power of artificial neural networks. An ANN can learn complex relationships from the input data using the parameters between the layers of the network. These parameters are optimized using backpropagation during the training phase. As the size and dimension of the input data grow, training a neural network with fewer parameters is a difficult task. Complex machine learning problems can be solved efficiently using a quantum computer [5]. Biamonte et.al [3] described the advantages of using quantum computing in solving machine learning problems. In their work [18, 19], Schuld et.al designed a quantum circuit as a classifier for binary classification using distance-based kernel function. Recent work [8, 10] states that machine learning problems can be solved efficiently using a hybrid classical-quantum approach that uses a classical computer to optimize quantum parameters. Tacchino et.al [22] implemented an artificial neuron on a quantum processor. Open-source software [6] are developed to implement quantum algorithms on a real quantum processor or a quantum simulator. Classical computing can store and process the data in the form of binary digits. In contrast, quantum computers store and process the data using qubits [13] which use both 0 and 1 at the same time.

1- Introduction

2- Related work

3- Optimization by continuation

4- Results and discussion

5- Conclusions and recommendations

References

Abstract

Artificial Neural Networks research field is among the areas of major activity in Artificial Intelligence. Training a neural network is an NP-hard optimization problem that presents several theoretical and computational limitations. In optimization, continuation refers to an homotopy transformation of the fitness function that is used to obtain simpler versions of such fitness function and improve convergence. In this paper we propose an approach for Artificial Neural Network training based on optimization by continuation and meta-heuristic algorithms. The goal is to reduce overall execution time of training without causing negative effects in accuracy. We use continuation together with Particle Swarm Optimization, Firefly Algorithm and Cuckoo Search for training neural networks on public benchmark datasets. The continuation variations of the studied meta-heuristic algorithms reduce execution time required to complete training in about 5–30% without statistically significant loss of accuracy when compared with standard variations of the meta-heuristics.

Introduction

In the latest years, we have seen the efforts in optimization field to solve very complex and challenging tasks. At the same time, the success in this area has led researchers and practitioners to address much larger instances and more difficult classes of problems, specially in meta-heuristic area [1,2]. Training a neural network is an NP-hard optimization problem [3] that usually involves thousands and millions of parameters. Additionally, the fitness function of Artificial Neural Networks (ANNs) is highly non-convex [4] and presents a poor correspondence between its local and global structure [5]. Also, the number of local minima increases exponentially respect to the number of parameters of the network. In fact, such local minima appears to be saddle points instead of true minima [6] reducing the convergence speed of optimization, specially for gradient based training algorithms. In general, second-order methods have not obtained better results than first-order methods in this area [5]. These facts become even more problematic as we start using modern deep learning architectures. ANNs are widely used in several applications of image recognition [7], signal processing [8], speech [9] and several other fields Nevertheless, neural networks are playing a major role in the field of pattern recognition and learning representations from data. The most popular algorithm for training is Stochastic Gradient Descent (SGD) which is a gradient based first order1 method that has been proven to converge into local minima in the parameter space. In general, first and second order2 optimization methods require the calculation of derivatives and Hessians for which Automatic Reverse Mode Differentiation (ARMD) have to be used [10]. The parallel implementation of such methods faces important challenges due to the inherent sequential nature of ARMD. Recently, several meta-heuristic algorithms for optimization have been actively used in ANN training. Meta-heuristic algorithms have features to escape local minima and increase the probability of global convergence. In the literature there are reports of natureinspired meta-heuristics such as Cuckoo Search [11,12], Firefly Algorithm [13,14], Wolf Search Optimization [15,16], Particle Swarm Optimization (PSO) [17,18] and others.

I- Introduction

II- Design of Analog and Digital Hybrid MAC Circuit

III- Experimental Results

IV- Conclusions

REFERENCES

Abstract

Demand for high-performance hardware acceleration for machine learning applications is increasing rapidly. This paper presents a low power analog and digital hybrid MAC (Multiply and Accumulation) circuit for artificial neural networks. The proposed MAC circuit consists of an analog synapse unit, digital preprocessing and postprocessing unit for support of the parallel analog synapse cores. As the hybrid MAC circuit supports relatively low power and fast multiple MAC operations, it provides a good advantage in developing hardware accelerator for artificial neural networks.

INTRODUCTION

Deep Neural Network (DNN) technology is rapidly evolving on the basis of the Internet which enables acquiring large amounts of learning data and hardware technology capable of high-speed parallel processing. In recent years, various Deep Neural Network (DNN) frameworks have been proposed such as R-CNN, SSD and YOLO V2, etc. Since the convolution operation is composed of millions of multiplications and additions, a high-performance accelerator having a parallel structure is required for real-time image recognition. The purpose of this study is to design a new MAC circuit that can be efficiently used in DNN for an image recognition field which is one of the most popular application fields of artificial neural networks. This paper is organized as follows. Section II presents the proposed MAC algorithm and the architecture of our hybrid MAC circuit. Synthesis results are given in Section III. Section IV presents the conclusions.

I- Introduction

II- Classification of plant diseases

III- Classification algorithms

IV- Conclusion

References

Abstract

Plants are considered to be important as they are the source of energy supply to mankind. Plant diseases can affect the leaf any time between sowing and harvesting which leads to huge loss on the production of crop and economical value of market. Therefore, leaf disease detection plays a vital role in agricultural field. However, it requires huge manpower, more processing time and extensive knowledge about plant diseases. Hence, machine learning is applied to detect diseases in plant leaves as it analyzes the data from different aspects, and classifies it into one of the predefined set of classes. The morphological features and properties like color, intensity and dimensions of the plant leaves are taken into consideration for classification. This paper presents an overview on various types of plant diseases and different classification techniques in machine learning that are used for identifying diseases in different plant leaves.

INTRODUCTION

INDIA is a fast developing country and agriculture is the back bone for the country’s development in its early stage. However, agricultural field faces lots of hurdles including huge loss in the crop production. Plant leaf diseases are one of the important reasons for the loss in the production and plant leaf disease identification is also very difficult in agriculture field. Naked eye method is a traditional method of identifying the diseases which involves huge man power, inaccurate, time consuming and not applicable for larger fields. In addition, it is very expensive as it requires continuous monitoring by the experts. Hence, machine learning; a reliable prediction methodology is used for detecting various diseases of plant leaves caused by fungus, bacteria and virus. However, disease prediction using classification algorithms appears to be a difficult task as the accuracy varies for different input data. In this paper, several research contributions related to various plant leaf diseases detection using different classification algorithms are reviewed and compared.