Search

There has been a lot of research on edge-computing task offloading in deep reinforcement learning (DRL). Deep reinforcement learning is one of the important algorithms in the current AI field, but there is still room for improvement in the time cost and adaptive correction ability of the algorithm. This paper studies the application of DRL algorithms in edge-computing task offloading, and its key innovation is to propose an MADRLCO algorithm, which is based on the design idea of the Actor–Critic framework, uses the DNN model to act as an Actor, and can more accurately locate the initial decision through iterative training, and use the LSTM model to optimize the Critic, which can be more accurate. The optimal decision can be located in a short period of time. The main work of this paper is divided into three parts: (1) The AC algorithm of the Actor–Critic framework in DRL is proposed to be applied to edge-computing task offloading. (2) To address the weak generalization ability of the basic version of the Actor–Critic algorithm in multi-objective optimization, the sequential quantitative correction and adaptive correction parameter K method are used to optimize the Critic frame, thereby improving the generalization ability of the model in multi-objective decision-making and improving the rationality of decision-making results. (3) Aiming at the problem of large time cost in the critical framework of the model, a search algorithm for resource allocation-related parameters based on the time-series prediction method is proposed (time-series forecasting is a research branch of pattern recognition), which reduces the time overhead of the algorithm and improves the adaptive correction capability of the model. The algorithm in this paper can adapt to not only the time-varying network channel state, but also the time-varying number of device connections. Finally, it is proved by experiments that compared with the DRL calculation offloading algorithm based on DNN plus binary search, the MADRLCO algorithm reduces the model training time by 66.27%, and in the environment of the time-varying number of devices in the metasystem, the average model average. The standard calculation rate is 0.0403 higher than that of the current optimal algorithm. [ABSTRACT FROM AUTHOR]

Computer Vision (CV) has become an essential field in Artificial Intelligence applications. Object detection and recognition (ODR) is one of the fundamental tasks of computer vision implementations. However, developing an efficient ODR model is still a significant problem. The model's execution time and speed are the most critical features during the inference or detection and recognition process, which need to be improved using the latest object detection architectures. In this paper, the handwritten detection and recognition (HDR) model is developed based on previously known algorithms with their efficiency, such as Faster R-CNN and YOLOv4 in the first hand. On the other hand, two new models capable of detecting and recognizing handwritten digits using the latest ODR algorithm are proposed, one based on the latest YOLO family architecture (YOLOv5-HDR) with high speed and accuracy and the other using the transformers architecture (DETR). To the best of our knowledge, this is the first study to achieve a details comparison between YOLOv5 and transformers-based models in handwritten digit detection. Finally, the detailed performance analysis achieved by the paper proves that the YOLOv4-based model achieved the testing inference 13% faster than Faster R-CNN. However, the proposed YOLOv5-based model outperformed the YOLOv4 and the transformers-based one as it increased the testing execution time 25% faster than the YOLOv4, three times faster than the DETR model. A further adversarial attack test has been conducted to ensure the robust performance of the proposed model. Furthermore, numerical experiment results and their analyses demonstrate the robustness and effectiveness of the proposed YOLOv5-based model being the most stable for handwritten digit detection and recognition tasks. [ABSTRACT FROM AUTHOR]

This paper proposes a novel algorithm that combines symbolic execution and data flow testing to generate test cases satisfying multiple coverage criteria of critical software applications. The coverage criteria considered are data flow coverage as the primary criterion, software safety requirements, and equivalence partitioning as sub-criteria. black The characteristics of the subjects used for the study include high-precision floating-point computation and iterative programs. The work proposes an algorithm that aids the tester in automated test data generation, satisfying multiple coverage criteria for critical software. The algorithm adapts itself and selects different heuristics based on program characteristics. The algorithm has an intelligent agent as its decision support system to accomplish this adaptability. Intelligent agent uses the knowledge base to select different low-level heuristics based on the current state of the problem instance during each generation of genetic algorithm execution. The knowledge base mimics the expert's decision in choosing the appropriate heuristics. black The algorithm outperforms by accomplishing 100% data flow coverage for all subjects. In contrast, the simple genetic algorithm, random testing and a hyper-heuristic algorithm could accomplish a maximum of 83%, 67% and 76.7%, respectively, for the subject program with high complexity. black The proposed algorithm covers other criteria, namely equivalence partition coverage and software safety requirements, with fewer iterations. black The results reveal that test cases generated by the proposed algorithm are also effective in fault detection, with 87.2% of mutants killed when compared to a maximum of 76.4% of mutants killed for the complex subject with test cases of other methods. [ABSTRACT FROM AUTHOR]

The accuracy and reliability of continuous space curve estimation is the key to global exploration. An improved artificial intelligence algorithm is proposed for continuous space analysis. First, a wavelet basis ANN algorithm is proposed to determine the discretization strategy in continuous space. The hidden layer node transfer function in a BP neural network is replaced by a wavelet basis function, and the modified BP neural network is composed of a wavelet neural network. Second, an improved wolf algorithm is established. The core wolf system ensures the precision of the whole exploration task. Finally, main and auxiliary double cores and a five-class decision factor are used to establish a population classification model to solve the convergence of the algorithm. [ABSTRACT FROM AUTHOR]

The modeling of channel and temporal information is of crucial importance for action recognition tasks. To build a high-performance action recognition network by effectively capturing channel and temporal information, we propose CLS-Net: an action recognition algorithm based on channel-temporal information modeling. The proposed CLS-Net characterizes channel and temporal information by inserting multiple modules to an end-to-end backbone network, including a channel attention module (CA module) for modeling channel information, a long-term temporal module (LT module) and a short-term temporal module (ST module) for modeling temporal information. Specifically, the CA module extracts the correlation between feature channels so the network can learn to selectively strengthen the features containing useful information and suppress the useless features through global information. The LT module moves some channels in the temporal dimension to realize information interaction across time domains and model global temporal information. The ST module enhances the motion-sensitive features by calculating the feature-level frame difference information and realizes the representation of local motion information. Since the multi-module insertion mode directly affects the whole model's final performance, we propose a novel multi-module insertion mode instead of a simple series or parallel connection to ensure that the multiple modules can complement one another and cooperate with each other more efficiently. CLS-Net achieves SOTA performance on the EgoGesture and Jester dataset in the same type of network and achieves competitive results on the Something-Something V2 dataset. [ABSTRACT FROM AUTHOR]

With the wide application of data mining and deep learning in mobile cellular network operation and maintenance, network measurement report (MR) plays an increasingly important role in artificial intelligence for IT operations (AIOps). For the integrity of MR reported by the operation and maintenance (OM) proxy of base station, existing collecting methods are typically based on static distributed clustering. Due to the lack of effective load balancing scheme, nevertheless, these methods typically result in some issues, e.g., low collecting efficiency, poor scalability, and excessive number of servers. Thus, in this work, leveraging the historical law of uploading MR for load forecasting, we propose the weighted least-connection load balancing algorithm (LPWLC) based on load forecasting. First, the historical law of reported MR is utilized to predict the load. Second, using the strategy of static binding and dynamic load adjustment, we bind OM with the assigned server in one cycle, calculate the server load in real-time, and evaluate the server weight by the load of each server. Finally, real-time load adjustment is carried out in line with the number of request connections and the weight of servers. Compared with the existing ones, the proposed algorithm could remove backup servers, thereby effectively reducing the cost and power consumption. Compared with the existing methods, this method has improved the load balancing degree by 28%, and reduced the energy consumption by 104 W per hour. [ABSTRACT FROM AUTHOR]

In order to succeed in our everyday life, efficient performance of the power system is of utmost importance, and hence, all the sub-sectors of the power system should be analyzed for the purpose of achieving efficiency and accuracy. It must also be remembered that load forecasting assists a lot to improve the power system. Moreover, it contributes substantially to formulate logical approaches for emerging short-term load forecasting (STLF) for all days including the distinct days and make them follow a uniform standard. Of all the techniques which have been applied so far, honey bee-optimized Euclidean norm, based on fuzzy inference system, is used for identifying the problems, and in addition, support vector classifier is utilized to prepare the STLF models. Parameters — temperature, humidity, monsoons, wind, cloud density, dew point, season, hour of the day, day of the week, distinct day, and holiday — have been taken into account for the current study. A well-prepared database can be used for regression which will be of immense help to forecast the load using artificial intelligence. For every day of a month, the MAPE is computed (using the forecasted and actual hourly values) in order to observe the accuracy of STLF. The planned method has been very successful for the load forecast of all days for all seasons. The forecast has been done using the technique for a real time data of one year (test forecast year) with a historical dataset collected for a period of two years, and the results obtained for all seasons have been found to be satisfactory. STLF has helped to find better values due to its pace, and become healthier than other methods already in practice. With the advent of smart grid, the data will be accessible at more granular level as smart meters have capability to provide consumer load, usage data on-line and this facility will be of great help to utility operators and planners for operations on-line. How to use the data available from smart meters for better STLF is a challenging task and it would draw much attention for future research. [ABSTRACT FROM AUTHOR]

One of the major challenges facing Artificial Intelligence in the future is the design of trustworthy algorithms. The development of trustworthy algorithms will be a key challenge in Artificial Intelligence for years to come. Cultural Algorithms (CAs) are viewed as one framework that can be employed to produce a trustable evolutionary algorithm. They contain features to support both sustainable and explainable computation that satisfy requirements for trustworthy algorithms proposed by Cox [Nine experts on the single biggest obstacle facing AI and algorithms in the next five years, Emerging Tech Brew, January 22, 2021]. Here, two different configurations of CAs are described and compared in terms of their ability to support sustainable solutions over the complete range of dynamic environments, from static to linear to nonlinear and finally chaotic. The Wisdom of the Crowds method was selected for the one configuration since it has been observed to work in both simple and complex environments and requires little long-term memory. The Common Value Auction (CVA) configuration was selected to represent those mechanisms that were more data centric and required more long-term memory content. Both approaches were found to provide sustainable performance across all the dynamic environments tested from static to chaotic. Based upon the information collected in the Belief Space, they produced this behavior in different ways. First, the topologies that they employed differed in terms of the "in degree" for different complexities. The CVA approach tended to favor reduced "indegree/outdegree", while the WM exhibited a higher indegree/outdegree in the best topology for a given environment. These differences reflected the fact the CVA had more information available for the agents about the network in the Belief Space, whereas the agents in the WM had access to less available knowledge. It therefore needed to spread the knowledge that it currently had more widely throughout the population. [ABSTRACT FROM AUTHOR]

In Internet of Things (IoT) business applications, healthcare records and medical web service management have emerging technologies tremendous amount of data in daily transactions. The cloud service providers have attracted the attention of a huge number of user requests based on different Quality of Service (QoS) factors. The data gained researchers attention to predict user behavior through the production of IoT applications. Artificial Intelligence (AI)-based techniques have a great impact on analyzing and detection of web service segmentation and user behavior in selecting and allocating online services and online stores with wireless communications and smart devices. This research improves the Redundancy, Frequency, and Maintenance value (RFM) model to evaluate healthcare records using a hybrid Grouping Genetic Algorithm (GGA) and Density-based Spatial Clustering of Applications with Noise (DBSCAN) algorithms. In this hybrid method, using the GGA algorithm, the features selection is applied to find the optimal features for healthcare records of customer segmentation. After that, by applying the RFM model on the data output of the genetic algorithm, the data are clustered. Finally, by applying the DBSCAN algorithm, the most suitable case for clustering will be selected. The simulation results show that the accuracy of the genetic algorithm is 97% and the final clustering accuracy is 92%. [ABSTRACT FROM AUTHOR]