Decision Module Research Articles

SiamATL: Online Update of Siamese Tracking Network via Attentional Transfer Learning.

Visual object tracking with semantic deep features has recently attracted much attention in computer vision. Especially, Siamese trackers, which aim to learn a decision making-based similarity evaluation, are widely utilized in the tracking community. However, the online updating of the Siamese fashion is still a tricky issue due to the limitation, which is a tradeoff between model adaption and degradation. To address such an issue, in this article, we propose a novel attentional transfer learning-based Siamese network (SiamATL), which fully exploits the previous knowledge to inspire the current tracker learning in the decision-making module. First, we explicitly model the template and surroundings by using an attentional online update strategy to avoid template pollution. Then, we introduce an instance-transfer discriminative correlation filter (ITDCF) to enhance the distinguishing ability of the tracker. Finally, we suggest a mutual compensation mechanism that integrates cross-correlation matching and ITDCF detection into the decision-making subnetwork to achieve online tracking. Comprehensive experiments demonstrate that our approach outperforms state-of-the-art tracking algorithms on multiple large-scale tracking datasets.

Allocation strategy for time‐sensitive tasks in mobile edge computing: An observable perspective

SummaryWith the development of computing technology and the popularization of application in 5G network, mobile‐edge computing (MEC), which can effectively reduce time delay and save energy consumption, has attracted extensive attention from the academic community. How to make a suitable allocation decision for tasks becomes one of the critical issues in MEC systems. In order to yield the most benefit to a newly arriving task, by placing a decision making module (DMM) in a mobile device, a MEC system architecture is presented. Based on the net benefit of a newly arriving task, the DMM makes a decision of dropping the task, allocating the task to the local execution system, or offloading the task to the MEC server via the transmission system. From the observable perspective of task individuals, a pure threshold strategy is proposed to show that a Nash equilibrium is always allowed. By constructing a two‐dimensional continuous‐time Markov chain, a social optimal threshold strategy is proposed. Numerical results show that threshold under the pure threshold strategy is always greater than that under the social optimal threshold strategy. For this, a charging policy is presented to coincide the pure threshold strategy with the social optimal threshold strategy.

Using Reinforcement Learning in a Layered Airspace to Improve Layer Change Decision

Current predictions for future operations with drones estimate traffic densities orders of magnitude higher than any observed in manned aviation. Such densities call for further research and innovation, in particular, into conflict detection and resolution without the need for human intervention. The layered airspace concept, where aircraft are separated per vertical layer according to their heading, has been widely researched and proven to increase traffic capacity. However, aircraft traversing between layers do not benefit from this separation and alignment effect. As a result, interactions between climbing/descending and cruising aircraft can lead to a large increase in conflicts and intrusions. This paper looks into ways of reducing the impact of vertical transitions within the environment. We test two reinforcement learning methods: a decision-making module and a control execution module. The former issues a lane change command based on the planned route. The latter performs operational control to coordinate the longitude and vertical movement of the aircraft for a safe merging manoeuvre. The results show that reinforcement learning is capable of optimising an efficient driving policy for layer change manoeuvres, decreasing the number of conflicts and losses of minimum separation compared to manually defined navigation rules.

AIMM: Artificial Intelligence Merged Methods for flood DDoS attacks detection

DDoS is one of the most common attacks on the web today. Hence, a quick detection system can enable automatic blocking or notification of an attack. In this paper, we propose a framework called AIMM (Artificial Intelligence Merged Methods). Our solution is based on three modules: preprocessing data incoming to the server, classification, and decision-making. The last stage is the decision-making module which gets the probability from all implemented AI methods and analyzes/aggregates them for making a final decision about the attack. The idea is based on the analysis of the TCP/UDP information reaching the target server and a quick decision method. The described technique is not limited to the selected AI method, and just for the tests, we used two different: neural networks and the k-nearest neighbors. As the aggregation solution, we used soft sets inference and averaging, weighted averaging technique. The proposal was subjected to performance tests on a publicly available database known as BOUN DDoS Dataset (and reached accuracy on 99,5%). The results were compared with the state-of-art and discussed in terms of its advantages and disadvantages.

Content features and its implementation in novel app-based psycho-oncological interventions for cancer survivors: a narrative review.

The interest in and availability of app-based psycho-oncological interventions for cancer survivors has grown rapidly. In this narrative review, we therefore aim to provide an overview of the psychological content features included in psycho-oncological apps and how the content is implemented in the apps. A total of 19 studies (six randomized controlled trials, six studies with prepost design, five study protocols and two app developments), all published in the last 18 months (July 2020 to December 2021), were analyzed in terms of the described app-based interventions and their psycho-oncological content. Apps for cancer survivors are conducted for a variety of different diagnoses and often target a specific patient population, for example patients with a specific therapy. The most frequent content features are psychoeducation, self-monitoring of symptoms, and some sort of exercise or relaxation program. Depending on the app's purpose, some offer specific content, such as decision aids, medication schedules or modules to improve coping strategies. Most apps combine informative and interactive elements. Psycho-oncological apps for cancer survivors target many relevant psychological problem areas and aim to improve patient empowerment. Available mHealth interventions can therefore be seen as a good complement for cancer care and also offer the possibility to improve patient involvement.

Actor-critic reinforcement learning for bidding in bilateral negotiation

Designing an effective and intelligent bidding strategy is one of the most compelling research challenges in automated negotiation, where software agents negotiate with each other to find a mutual agreement when there is a conflict of interests. Instead of designing a hand-crafted decision-making module, this work proposes a novel bidding strategy adopting an actor-critic reinforcement learning approach, which learns what to offer in a bilateral negotiation. An entropy reinforcement learning framework called Soft Actor-Critic (SAC) is applied to the bidding problem, and a self-play approach is employed to train the model. Our model learns to produce the target utility of the coming offer based on previous offer exchanges and remaining time. Furthermore, an imitation learning approach called behavior cloning is adopted to speed up the learning process. Also, a novel reward function is introduced that does take not only the agent?s own utility but also the opponent?s utility at the end of the negotiation. The developed agent is empirically evaluated. Thus, a large number of negotiation sessions are run against a variety of opponents selected in different domains varying in size and opposition. The agent?s performance is compared with its opponents and the performance of the baseline agents negotiating with the same opponents. The empirical results show that our agent successfully negotiates against challenging opponents in different negotiation scenarios without requiring any former information about the opponent or domain in advance. Furthermore, it achieves better results than the baseline agents regarding the received utility at the end of the successful negotiations.

Real-Time Safety Decision-Making Method for Multirotor Flight Strategies Based on TOPSIS Model

Multirotors play an important role in electric power inspection, border control, modern agriculture, forest fire fighting, flood control, disaster prevention, etc. Multirotor failures, such as a communication fault, a sensor failure, or a power system anomaly, may well lead to mission interruption, multirotor crashes, and even casualties. To ensure flight safety, a multirotor decision module should be established to prevent or reduce the adverse effects caused by failure. Therefore, this paper proposes a real-time safety decision-making method for multirotor flight strategies based on the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS). Firstly, the flight of the multirotor was simulated based on the Rflysim UAV flight simulation platform, and a fault-injection module was constructed to simulate different types of faults, so as to realize real-time monitoring of the flight status of the multirotor, and to collect flight data under various faults to establish condition assessment information sources. Then, based on the random forest algorithm, a failure level classification model of the multirotor was constructed, the model was trained and verified by inputting flight data of three types of safety level failures, and the model effectively classified the failure levels of the multirotor. Under this framework, a real-time safety decision-making model for the multirotor based on the TOPSIS model was constructed to realize the flight safety decision-making of the multirotor under different faults. This method can effectively realize the real-time decision-making for the flight strategy of a multirotor. By comparison with other models, the classification accuracy of the failure level classification model is higher, and the consideration of flight decision-making is more comprehensive and accurate, thus effectively ensuring the flight safety of the multirotor.

Runtime Assurance of Learning-Based Lane Changing Control for Autonomous Driving Vehicles

Learning techniques such as deep reinforcement learning have been increasingly used in the controller design for autonomous vehicles, e.g., lane changing controllers. Although the use of learning techniques can remarkably increase the level of autonomy, their presence poses a great challenge for safety assurance due to the black-box and data-driven nature of learning techniques. In this work, we study the problem of how to safeguard the learning-based lane changing controller for collision avoidance. Our solution leverages on the runtime assurance framework, in particular the Simplex architecture to bound the behavior of the learning-based controller and to provide safety guarantees. The basic idea is to encompass the learning-based controller with a safety-by-construction controller and a decision module, which monitors the output of the learning-based controller at runtime and implements a switching logic between these two controllers according to the changing environment. We present the detailed design of the decision module and formally prove its correctness for collision avoidance. We also carry out a comprehensive experimental evaluation in a set of realistic highway scenarios using the SUMO simulator. The simulation results show that our proposed solution can not only provide safety guarantee for the learning-based lane changing controller, but also maintain a considerable level of efficiency in different volumes of traffic flow.

Implementation of the internet of things network for monitoring audio information on a microprocessor and controller

The subject of research is the development and implementation of Internet of Things (IoT) network structures for monitoring and analyzing audio information based on Raspberry microprocessor (MP) and an Arduino controller. The purpose of the article is to detail the process of developing an IoT based audio information monitoring network and evaluate the results. The authors have developed two variants of IoT structures for monitoring and analyzing audio and voice information. The IoT network includes a sound sensor (microphone), a unit for analyzing the information received from it and a decision-making module. A diagram of the first IoT structure for assessing the sound level based on the MP and controller is given.The algorithm of IoT network functioning for the analysis of voice information is detailed. It includes receiving information from the microphone, transmitting this information to the MP, processing it according to certain rules, forming a solution by the controller and issuing recommendations to a user. The algorithm is implemented in the IoT network, which includes a microphone, Raspberry MP, Arduino controller, software, applications for the operator.A prototype of the IoT network was created for the analysis of voice information and experiments were conducted to test its functioning. The audio recognizer was trained using various audio samples. The voice sound analysis system was tested in four scenarios: with a large and small amount of background noise, loud and quiet voice. Analysis of the results of the experiment showed that the voice sound analysis system works better when the voice is loud, as well as in a place where the situation is with minimal background noise.

IoT-Inspired Monitoring Framework for Real-Time Stereotypic Movement Analysis

Physical activity recognition has become one of the prominent domain of research in the field of smart healthcare. The identification of physical activities through wearable sensors provide significant information about the functional ability and lifestyle of an individual. In this article, a smart physical activity-assisted behavior prediction framework is presented to recognize different stereotypical motor movements of the individual suffering from autism spectrum disorder in real time. This article employs the data acquisition efficiency of Internet of Things (IoT) assisted smart wearable sensors to capture the physical activities of an individual. To predict the physical activities, a deep learning-inspired multilayered convolutional neural network (CNN) + long short-term memory network is proposed. Furthermore, the utility of the framework is increased by presenting an effective fog-assisted alert-based decision-making module. The decision-making module notifies the concerned caretaker or medical specialist about the performed irregular physical activity in real time. Moreover, the process of cloud-based record generation also enhances the significance of the proposed solution. The proposed approach is validated by comparing the prediction performance with traditional handcraft-based and modern deep learning-based state-of-the-art approaches such as <formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex>$k$</tex></formula> -nearest neighbors, hidden Markov model, and CNN, where it outperformed by achieving the overall accuracy of 91.88%.

Adaptive 3D routing protocol for flying ad hoc networks based on prediction-driven Q-learning

The routing protocols are paramount to guarantee the Quality of Service (QoS) for Flying Ad Hoc Networks (FANETs). However, they still face several challenges owing to high mobility and dynamic topology. This paper mainly focuses on the adaptive routing protocol and proposes a Three Dimensional Q-Learning (3DQ) based routing protocol to guarantee the packet delivery ratio and improve the QoS. In 3DQ routing, we propose a Q-Learning based routing decision scheme, which contains a link-state prediction module and routing decision module. The link-state prediction module allows each Unmanned Aerial Vehicle (UAV) to predict the link-state of Neighboring UAVs (NUs), considering their Three Dimensional mobility and packet arrival. Then, UAV can produce routing decisions with the help of the routing decision module considering the link-state. We evaluate the various performance of 3DQ routing, and simulation results demonstrate that 3DQ can improve packet delivery ratio, goodput and delay of baseline protocol at most 71.36%, 89.32% and 83.54% in FANETs over a variety of communication scenarios.

A new occluded face recognition framework with combination of both Deocclusion and feature filtering methods.

Face recognition plays the significant role in many human-computer interaction decvices and applications, whose access control systems are based on the verification of face biometrical features. Though great improvement in the recognition performances have been achieved, when under some specific conditions like faces with occlusions, the performance would suffer a severe drop. Occlusion is one of the most significant reasons for the performance degrade of the existing general face recognition systems. The biggest problem in occluded face recognition (OFR) lies in the lack of the occluded face data. To mitigate this problem, this paper has proposed one new OFR network DOMG-OFR (Dynamic Occlusion Mask Generator based Occluded Face Recognition), which keeps trying to generate the most informative occluded face training samples on feature level dynamically, in this way, the recognition model would always be fed with the most valuable training samples so as to save the labor in preparing the synthetic data while simultaneously improving the training efficiency. Besides, this paper also proposes one new module called Decision Module (DM) in an attempt to combine both the merits of the two mainstream methodologies in OFR which are face image reconstruction based methodologies and the face feature filtering based methodologies. Furthermore, to enable the existing face deocclusion methods that mostly target at near frontal faces to work well on faces under large poses, one head pose aware deocclusion pipeline based on the Condition Generative Adversarial Network (CGAN) is proposed. In the experimental parts, we have also investigated the effects of the occlusions upon face recognition performance, and the validity and the efficiency of our proposed Decision based OFR pipeline has been fully proved. Through comparing both the verification and the recognition performance upon both the real occluded face datasets and the synthetic occluded face datasets with other existing works, our proposed OFR architecture has demonstrated obvious advantages over other works.

Performance investigation of the probability-aided maximum likelihood sequence detector for a probabilistic shaped 16-QAM system.

In this paper, for the first time, a probability-aided maximum-likelihood sequence detector (PMLSD) is experimentally investigated through a 64-GBaud probabilistic shaped 16-ary quadrature amplitude modulation (PS-16QAM) transmission experiment. In order to relax the impacts of PS technology on the decision module, a PMLSD decision scheme is investigated by modifying the decision criterion of maximum-likelihood sequence detector (MLSD) correctly. Meanwhile, a symbol-wise probability-aided maximum a posteriori probability (PMAP) scheme is also demonstrated for comparison. The results show that the PMLSD scheme outperforms the direct decision scheme about 1.0-dB optical signal to noise ratio (OSNR) sensitivity. Compared with symbol-wise PMAP scheme, PMLSD scheme can effectively relax the impacts of PS technology on the decision module and a more than 0.8-dB improvement in terms of OSNR sensitivity in back-to-back (B2B) case is obtained. Finally, we successfully transmit the PS-16QAM signals over a 2400-km fiber link with a bit error ratio (BER) lower than 1.00×10-3 by adopting the PMLSD scheme.

LLNet: A Fusion Classification Network for Land Localization in Real-World Scenarios

Lane localization is one of the core tasks in an autonomous driving system. It receives the visual information collected by the camera and the lane marks and road edges information outputted from the perception module and gives lane index for the subsequent decision module. Traditional rule-based lane localization methods using navigation maps can only be effective in regular road scenarios and have poor generalization ability. High-Definition Map (HD map) was originally thought to solve the lane localization problem, but due to the regulations of the relevant departments, HD map is currently not allowed to be used in autonomous driving systems. In addition, many multi-sensor fusion methods have been proposed to solve the lane localization problem. However, due to the extremely strict safety requirements of autonomous driving systems, these well-designed solutions make it difficult to meet the requirements in terms of robustness, efficiency, and stability. To solve these problems, we innovatively define the lane localization task as a classification problem. First, to better utilize the perceptual information outputted from the perceptual model, we design an image-generating method that projects the perceptual information onto a new image and ensures that our model can learn the perceptual features wisely. Second, to better fuse the perceptual and visual information, we propose a fusion structure deep learning neural network named LLNet to address the lane localization problem in an end-to-end manner. Finally, to ensure the generalization ability, robustness, and stability of LLNet, we conduct extensive comparison experiments on a large-scale real-world dataset, with a total mileage of over 400 km. The experiments show that our approach remarkably outperforms the deep learning classification baselines. In the discussion part of this paper, we give a comprehensive and detailed elaboration for the effectiveness of various designs in our LLNet. To our knowledge, LLNet is the first lane localization method based entirely on deep learning. LLNet is added to the self-driving suite for a kind of mass production vehicle that will be available in the summer of 2022, with an expected sales volume more than 300,000.

Exploiting Diverse Information in Pre-Trained Language Model for Multi-Choice Machine Reading Comprehension

Answering different multi-choice machine reading comprehension (MRC) questions generally requires different information due to the abundant diversity of the questions, options and passages. Recently, pre-trained language models which provide rich information have been widely used to address MRC tasks. Most of the existing work only focuses on the output representation at the top layer of the models; the subtle and beneficial information provided by the intermediate layers is ignored. This paper therefore proposes a multi-decision based transformer model that builds multiple decision modules by utilizing the outputs at different layers to confront the various questions and passages. To avoid the information diversity in different layers being damaged during fine-tuning, we also propose a learning rate decaying method to control the updating speed of the parameters in different blocks. Experimental results on multiple publicly available datasets show that our model can answer different questions by utilizing the representation in different layers and speed up the inference procedure with considerable accuracy.

Altruism under Stress: Cortisol Negatively Predicts Charitable Giving and Neural Value Representations Depending on Mentalizing Capacity.

Altruism, defined as costly other-regarding behavior, varies considerably across people and contexts. One prominent context in which people frequently must decide on how to socially act is under stress. How does stress affect altruistic decision-making and through which neurocognitive mechanisms? To address these questions, we assessed neural activity associated with charitable giving under stress. Human participants (males and females) completed a charitable donation task before and after they underwent either a psychosocial stressor or a control manipulation, while their brain activity was measured using functional magnetic resonance imaging. As the ability to infer other people's mental states (i.e., mentalizing) predicts prosocial giving and may be susceptible to stress, we examined whether stress effects on altruism depend on participants' general capacity to mentalize, as assessed in an independent task. Although our stress manipulation per se had no influence on charitable giving, increases in the stress hormone cortisol were associated with reductions in donations in participants with high mentalizing capacity, but not in low mentalizers. Multivariate neural response patterns in the right dorsolateral prefrontal cortex (DLPFC) were less predictive of postmanipulation donations in high mentalizers with increased cortisol, indicating decreased value coding, and this effect mediated the (moderated) association between cortisol increases and reduced donations. Our findings provide novel insights into the modulation of altruistic decision-making by suggesting an impact of the stress hormone cortisol on mentalizing-related neurocognitive processes, which in turn results in decreased altruism. The DLPFC appears to play a key role in mediating this cortisol-related shift in altruism.SIGNIFICANCE STATEMENT Altruism is a fundamental building block of our society. Emerging evidence indicates a major role of acute stress and stress-related neuromodulators in social behavior and decision-making. How and through which mechanisms stress may impact altruism remains elusive. We observed that the stress hormone cortisol was linked to diminished altruistic behavior. This effect was mediated by reduced value representations in the right dorsolateral prefrontal cortex and critically depended on the individual capacity to infer mental states of others. Our findings provide novel insights into the modulation of human altruism linked to stress-hormone dynamics and into the involved sociocognitive and neural mechanisms, with important implications for future developments of more targeted interventions for stress-related decrements in social behavior and social cognition.

A visual approach towards forward collision warning for autonomous vehicles on Malaysian public roads.

Background: Autonomous vehicles are important in smart transportation. Although exciting progress has been made, it remains challenging to design a safety mechanism for autonomous vehicles despite uncertainties and obstacles that occur dynamically on the road. Collision detection and avoidance are indispensable for a reliable decision-making module in autonomous driving. Methods: This study presents a robust approach for forward collision warning using vision data for autonomous vehicles on Malaysian public roads. The proposed architecture combines environment perception and lane localization to define a safe driving region for the ego vehicle. If potential risks are detected in the safe driving region, a warning will be triggered. The early warning is important to help avoid rear-end collision. Besides, an adaptive lane localization method that considers geometrical structure of the road is presented to deal with different road types. Results: Precision scores of mean average precision (mAP) 0.5, mAP 0.95 and recall of 0.14, 0.06979 and 0.6356 were found in this study. Conclusions: Experimental results have validated the effectiveness of the proposed approach under different lighting and environmental conditions.

ГИБРИДНАЯ МЕТОДИКА ПРАКТИЧЕСКОЙ РЕАЛИЗАЦИИ СИСТЕМЫ ПРИНЯТИЯ РЕШЕНИЙ ПО ПРИОРИТЕТНОМУ РЕГУЛИРОВАНИЮ

The task of building a decision-making system within the framework of automated systems ofaccident-free control of technological processes based on forecasting models is considered.The analysis of models and methods of emergency forecasting is presented. The task of developinga methodology for practical implementation of the system based on the integral criterion of accident-free control, taking into account the time reserves to bring the process to a normal state(emergency forecasting system) and the resource component (preventive maintenance system) isformulated. The conclusion is made about the expediency of building decision-making systems andautomated control systems based on forecasting models, as the most promising approach to solvingthe problem of accident-free control of technological processes. The principle of building adecision-making system is based on the use of the integral criterion of accident-free management.The block diagram of the algorithm for calculating the integral criterion of accident-free control ispresented. Hybrid methodology for practical realization of such systems on the basis of priorityregulation, which includes a standard regulator, is offered. The procedure of formation of priorityregulators according to the forecast data is described. A block diagram of the algorithm of thepriority regulator, which determines the critical parameter on the basis of sensitivity theory, ispresented. In case of a positive forecast on an accident, the critical parameter is selected by themaximum of the sensitivity coefficient, and the minimum or maximum value of the parameter dependingon the sign of its rate of change is fed to the standard regulator from the matrix of criticalvalues as a set point. The structure of the decision-making system based on the concept of accident-free control of technological processes is given. The station of accident-free control forms thedata for the decision-making module on the basis of the compositional model of emergency forecastingand the integral criterion of accident-free control. Algorithm block diagram of the decisionmaking module for priority regulation is given.

UFS-Net: A unified flame and smoke detection method for early detection of fire in video surveillance applications using CNNs

Fire is a recurring event that usually causes a lot of social, environmental, ecological, and economic damage in different environments. Therefore, machine vision-based fire detection can be one of the most important tasks in modern surveillance systems. Most of the existing computer vision-based fire detection methods are only able to detect a single case of flame or smoke. In this research, a unified flame and smoke detection approach, termed “UFS-Net,” based on deep learning is proposed. An efficient and tailored convolutional neural network architecture is designed to detect both fire flames and smoke in video frames. UFS-Net is capable of identifying fire hazards by classifying video frames into eight classes: 1) flame, 2) white smoke, 3) black smoke, 4) flame and white smoke, 5) flame and black smoke, 6) black smoke and white smoke, 7) flame, white smoke and black smoke, and 8) normal status. To further increase the reliability of UFS-Net, a decision module based on a voting scheme is applied. In addition, a rich annotated dataset named “UFS-Data” that includes 849,640 images and 26 videos, captured/collected from various data sources and artificial images made in this research, is prepared for training and evaluation of UFS-Net. Extensive experiments conducted on “UFS-Data” and other benchmark datasets (i.e., “Mivia,” “BoWFire,” and “FireNet”), and the comparisons with state-of-the-art methods, confirm the high performance of UFS-Net. All the implementation source codes and the “UFS-Data” are made publicly available at https://github.com/alihosseinice/UFS-Net.

A Dynamic Service Placement Based on Deep Reinforcement Learning in Mobile Edge Computing

Mobile edge computing is an emerging paradigm that supplies computation, storage, and networking resources between end devices and traditional cloud data centers. With increased investment of resources, users demand a higher quality-of-service (QoS). However, it is nontrivial to maintain service performance under the erratic activities of end-users. In this paper, we focus on the service placement problem under the continuous provisioning scenario in mobile edge computing for multiple mobile users. We propose a novel dynamic placement framework based on deep reinforcement learning (DSP-DRL) to optimize the total delay without overwhelming the constraints on physical resources and operational costs. In the learning framework, we propose a new migration conflicting resolution mechanism to avoid the invalid state in the decision module. We first formulate the service placement under the migration confliction into a mixed-integer linear programming (MILP) problem. Then, we propose a new migration conflict resolution mechanism to avoid the invalid state and approximate the policy in the decision modular according to the introduced migration feasibility factor. Extensive evaluations demonstrate that the proposed dynamic service placement framework outperforms baselines in terms of efficiency and overall latency.