Output Decisions Research Articles

Explainable AI for Medical Data: Current Methods, Limitations, and Future Directions

With the power of parallel processing, large datasets,and fast computational resources, deep neural networks (DNNs) have outperformed highly trained and experienced human experts in medical applications. However, the large global community of healthcare professionals, many of whom routinely face potentially life-or-death outcomes with complex medicolegal consequences, have yet to embrace this powerful technology. The major problem is that most current AI solutions function as a metaphorical black-box positioned between input data and output decisions without a rigorous explanation for their internal processes. With the goal of enhancing trust and improving acceptance of AI-based technology in clinical medicine, there is a large and growing effort to address this challenge using eXplainable AI (XAI), a set of techniques, strategies, and algorithms with an explicit focus on explaining the “hows and whys” of DNNs. Here, we provide a comprehensive review of the state-of-the-art XAI techniques concerning healthcare applications and discuss current challenges and future directions. We emphasize the strengths and limitations of each category, including image, tabular, and textual explanations, and explore a range of evaluation metrics for assessing the effectiveness of XAI solutions. Finally, we highlight promising opportunities for XAI research to enhance the acceptance of DNNs by the healthcare community.

Optimal decisions of closed-loop supply chain under government recycling subsidy and value co-creation

Several countries and regions have adopted recycling subsidy policies to encourage recycling activities and enacted recycling laws to improve the recycling rate of waste electrical and electronic equipment (WEEE). These initiatives are designed to reduce the damage to the environment and natural resource consumption caused the enterprise. In this study, we analyze how government recycling regulations, value co-creation activities, and recycling subsidies influence the pricing and output decisions of members of a closed-loop supply chain. We develop a Stackelberg game model led by manufacturers and followed by retailers to explore the impact of value co-creation activities and recycling subsidies on the quantity of new and remanufactured products, recyclers’ prices and prices of recycled electronic parts, manufacturers’ capacity to recycle, and environmental performance in the context of government recycling regulations. Manufacturers can choose from three recycling modes: mandatory minimum remanufacturing (SDWmodel), non-mandatory partial remanufacturing (SDOmodel), and full remanufacturing (SDBmodel). The results show that 1) higher recycling subsidies optimize the recycling capacity of manufacturers, increasing the number of remanufactured products and reducing the market share of new products, which affects the price of new products. 2) higher recycling subsidies reduce the price of recycled electronic parts purchased by manufacturers, strengthen the bargaining power of consumers, increase the price of recyclers, promote the active participation of consumers in value co-creation activities, and increase the level of value co-creation. 3) higher recycling subsidies help to increase the economic value of all parties in the closed-loop supply chain (manufacturers, recyclers, consumers), but stricter recycling constraints are not always beneficial to environmental performance; environmental taxes and fees above a certain threshold of recycling constraints reduce the negative impact of production activities on the environment. The novelties and contributions of this research include proposing a new application for the Value Co-creation and Government policy utilized to Supply Chain and their impact on sustainable supply chain. In addition, we proposed three new recycling strategies and analysed the impact of environmental tax on the environmental performance. Therefore, this study aims to provide some insights for government policymakers and for managers when facing a more complex business environment.

Sales‐based compensation and collusion with heterogeneous firms

AbstractPricing and output decisions are often delegated to managers compensated on the basis of sales. Prior literature has shown that when firms are homogeneous, the delegation of pricing or output decisions to managers, compensated on the basis of sales, does not facilitate collusion. We show that when firms are heterogeneous, either in marginal cost or product quality, sales‐based compensation can facilitate collusion under both price and quantity competition. As a result, compensating managers on the basis of sales can increase firm profits and reduce consumer welfare. Additionally, we find that owners can strategically design managerial compensation structures to incentivize collusion between rival managers.

A novel adaptive coordinated tracking control scheme for a morphing aircraft with telescopic wings

Inspired by flight biology, morphing flight technology has great potential to improve the adaptability and maneuverability of aircraft. This paper is devoted to the flight control problem of morphing aircraft, and aimed at safe and fuel-saving flight through morphing actively. Specifically, the longitudinal dynamics of a morphing aircraft with telescopic wings is modelled as a strict-feedback nonlinear system. Through fitting the expression of aerodynamic parameters by the morphing ratio, the model uncertainties induced by morphing errors are embedded in the dynamics. To meet the safety and fuel-saving requirements, an Adaptive Coordinated Tracking Control Scheme (ACTCS) is then proposed, which consists of a morphing control module and a tracking control module. For the morphing control module, an on-line morphing decision model is given in an optimization process with respect to the morphing ratio, and a second-order tracking filter is introduced to smooth the decision output and ensure the physical realizability. For the tracking control module, the novel adaptive controllers for the velocity and altitude subsystems are proposed based on the dynamic surface control method, in which adaptive mechanisms are designed to compensate for the model uncertainties. Finally, the proposed ACTCS is simulated in nine different cases of the test flight mission, to verify its effectiveness, robustness and fuel-saving effect.

Implementation of Decision-Making Systems Based on a Typical Decisive Element

As large amounts of data are received and accumulated, the need to create automated and automatic decision-making systems for a wide variety of tasks becomes more and more urgent. In the technical field, one of them is technical diagnostics, and in medicine, diagnostics of the human condition. Currently, dozens of computer systems for continuous vibration control and monitoring of complex rotary-type units are in commercial operation, which form time trends for 14 vibration parameters with a time step of 1–8 s for each control point at the operated facility. The functionality of a universal decision-making module is proposed, the input data of which are the parameters and characteristics of the observed object, with the required resulting output decision.

Supply Chain Stability and Output Decision Model Considering Supply Disruption

Supply Chain Stability and Output Decision Model Considering Supply Disruption

Credibility-Supervised Dynamic Clustering and Consensus Model Involving Minority Opinions Handling in Social-Network Large-Scale Decision Making

Existing studies on social-network large-scale decision making (SN-LSDM) have overlooked two vital roles of the information credibility of decision makers (DMs): 1) Information credibility directly determines the reliability of clustering results and final decision output; 2) Consideration of minority opinions may hold the truth but may also be risky, while the credibility analysis of minority opinions can help avoid risks. To this end, this paper proposes credibility-supervised dynamic clustering and consensus model involving minority opinion processing. Firstly, DMs' information is analyzed and preprocessed to obtain DMs' credibility evaluation. Subsequently, DMs' credibility is used as one of the clustering criteria to supervise and guide the trust relationships and the clustering of large-scale group, which assists in deriving the reliable clustering result. This clustering is dynamic because the clustering result may alter as the DMs' information changes in consensus. Afterwards, an SN-LSDM consensus model with minority opinions settlement is developed. A credibility-based mechanism for identification, DMs' discussion and weight adjustment of minority opinion subgroup is constructed, including it trust risk measurement method and trust risk level-based weight adjustment. Then, a risk index is defined in feedback adjustment mechanism to adjust DMs' information, and a decision plan with a higher consensus level is obtained. Finally, the feasibility and effectiveness of the proposed model are verified through the case study of “916” Lu County earthquake relief. Discussions and comparisons are conducted by simulation experiments to explore the capabilities and advantages of our proposal.

Human Activity Recognition Using Ensemble Neural Networks and The Analysis of Multi-Environment Sensor Data Within Smart Environments

The significant focus and potential value of Human Activity Recognition (HAR) technologies based on non-invasive ambient sensors have been attributed to the advancement of Artificial Intelligence (AI) and the widespread adoption of sensors. Due to the proactive engagement of human activities and the utilization of Machine Learning (ML) techniques that depend on domain expertise, developing a standardized model for comprehending the everyday actions of diverse individuals has significant challenges. A technique for recognizing the user's everyday activities in multi-tenant intelligent environments has been developed. This methodology considers data feature limits and recognition approaches and is designed to limit sensor noise during human activities. This work aims at enhancing the quality of a publicly accessible HAR dataset to facilitate data-driven HAR.Additionally, the paper proposes a novel ensemble of neural networks (NN) as a data-driven HAR classifier. A Spatial Proximity Matrix (SPM)uses ambient sensors to facilitate contextawareness and mitigate data noise. The proposed method, named Homogeneous Ensemble Neural Network and Multi-environment Sensor Data (HENN-MSD), leverages a combination of a homogeneous ensemble NN and multi-environment sensor data to identify what individuals do in daily life accurately. The study featured the generation and integration of four fundamental models using the support-function fusion approach. This method included the computation of an output decision score for each basis classifier. The analysis of a comparative experiment conducted on the CASAS dataset indicates that the proposed HENN-MSD technique exhibits superior performance compared to the state-of-the-art methods in terms of accuracy (96.57%) in HAR.

An intelligent encryption decision method for autonomous domain of multilayer satellite network

Multi-layer satellite network (MLSN) is an important part of the Space-Air-Ground Integrated Network (SAGIN), it has obvious advantages in coverage density, business processing, service efficiency and other comprehensive capabilities compared with ordinary single-layer satellite networks. However, due to the fact that many space nodes, highly dynamic topology changes and wireless links scattered in free space, MLSNs are facing the dual challenges of inefficient organization and lack of security. In this case, this paper considers providing secure encryption decision support for the return of data obtained by low earth orbit (LEO) observation satellites relying on the local autonomous region of multi-layer satellite space. By quantifying the security strength of encryption algorithms, the work to be optimized is described as the problem of maximizing the security strength under the constraint of return delay of the data. This paper thus proposes a solution based on deep reinforcement learning (DRL), which relies on reasonably setting the key supporting elements such as state, action and reward closely related to the problem to achieve fast and reliable decision output, and the simulation results show that the proposed method has better performance than many comparison methods.

Conditional Predictive Behavior Planning With Inverse Reinforcement Learning for Human-Like Autonomous Driving

Making safe and human-like decisions is an essential capability of autonomous driving systems, and learning-based behavior planning presents a promising pathway toward achieving this objective. Distinguished from existing learning-based methods that directly output decisions, this work introduces a predictive behavior planning framework that learns to predict and evaluate from human driving data. This framework consists of three components: a behavior generation module that produces a diverse set of candidate behaviors in the form of trajectory proposals, a conditional motion prediction network that predicts future trajectories of other agents based on each proposal, and a scoring module that evaluates the candidate plans using maximum entropy inverse reinforcement learning (IRL). We validate the proposed framework on a large-scale real-world urban driving dataset through comprehensive experiments. The results show that the conditional prediction model can predict distinct and reasonable future trajectories given different trajectory proposals and the IRL-based scoring module can select plans that are close to human driving. The proposed framework outperforms other baseline methods in terms of similarity to human driving trajectories. Additionally, we find that the conditional prediction model improves both prediction and planning performance compared to the non-conditional model. Lastly, we note that the learning of the scoring module is crucial for aligning the evaluations with human drivers.

Evaluating batteries for renewable energy storage: A hybrid MCDM framework based on combined objective weights and uncertainty-preserved COPRAS

Battery technologies offer promising solutions for renewable energy storage. However, selecting the most suitable battery requires proper investigation. This study introduces a multi-criteria decision-making framework for assessing batteries based on various criteria and uncertain data, by using a combined objective weighting method and an uncertainty-preserved complex proportional assessment (UP-COPRAS). The proposed weighting method ensures objectivity and fairness in the weighting result by integrating interval entropy and a gray relational coefficient-supported decision-making trial and evaluation laboratory to capture variation and correlation degrees among the criteria. After incorporating interval numbers with a compensatory ranking method, the UP-COPRAS prioritizes batteries in a simple yet rigorous way using uncertain evaluation data. To test the feasibility of the framework, an illustrative case was employed to assess four battery alternatives using a five-dimensional criteria system. Through results comparison, two mathematical contributions are confirmed. First, the combined objective weighting method uses the variation and correlation features of numerical data to determine criteria weights, which prevents subjective manipulation and eliminates bias in statistical analysis. Second, the UP-COPRAS preserves uncertainties throughout the evaluation, resulting in a rational decision output by eliminating interference in the original data.

Xaitk-Saliency: An Open Source Explainable AI Toolkit for Saliency

Advances in artificial intelligence (AI) using techniques such as deep learning have fueled the recent progress in fields such as computer vision. However, these algorithms are still often viewed as "black boxes", which cannot easily explain how they arrived at their final output decisions. Saliency maps are one commonly used form of explainable AI (XAI), which indicate the input features an algorithm paid attention to during its decision process. Here, we introduce the open source xaitk-saliency package, an XAI framework and toolkit for saliency. We demonstrate its modular and flexible nature by highlighting two example use cases for saliency maps: (1) object detection model comparison and (2) doppelganger saliency for person re-identification. We also show how the xaitk-saliency package can be paired with visualization tools to support the interactive exploration of saliency maps. Our results suggest that saliency maps may play a critical role in the verification and validation of AI models, ensuring their trusted use and deployment. The code is publicly available at: https://github.com/xaitk/xaitk-saliency.

Securing Lifelines: Safe Delivery of Critical Services in Areas with Volatile Security Situation via a Stackelberg Game Approach

Vaccine delivery in under-resourced locations with security risks is not just challenging but also life threatening. The COVID pandemic and the need to vaccinate added even more urgency to this issue. Motivated by this problem, we propose a general framework to set-up limited temporary (vaccination) centers that balance physical security and desired (vaccine) service coverage with limited resources. We set-up the problem as a Stackelberg game between the centers operator (defender) and an adversary, where the set of centers is not fixed a priori but is part of the decision output. This results in a mixed combinatorial and continuous optimization problem. As part of our scalable approximation solution, we provide a fundamental contribution by identifying general duality conditions of switching max and min when both discrete and continuous variables are involved. Via detailed experiments, we show that the solution proposed is scalable in practice.

Hierarchical Neural Network and Simulation Based Structural Defect Identification and Classification

A vibration data-driven structural defect identification and classification technique is developed using frequency response under random excitation and a hierarchical neural network. A system of artificial neural networks (ANNs) is trained using finite element simulation-based synthetic data to reduce the need for many sensor measurements required otherwise. Principal component analysis (PCA) is employed to compress the high dimensionality of the vibration response data and eliminate the noise effect in the training and testing. Frequency responses data dimension for the structure with defects such a crack from stress concentration, rivet hole expansion, and attached foreign object mass such as ice accumulation in aircraft wing or fuselage are reduced using PCA and fed to a classifier network. The probabilistic decision output from the classifier network and the compressed data are then fed to the next levels of estimator networks, where each network is dedicated to the individual type of defect for the estimation of the defect parameters corresponding to that class of defect. The methodology is applied to a stiffened panel structure. The cracks and rivet hole expansions are introduced in the rivet line of the stiffener, and the foreign object mass is attached to the panel surface. The results show that it is possible to classify the defects and further estimate the defect parameters with good accuracy and reliability. It was observed that the damage classification network had an accuracy of roughly 95%. The damage localization network for crack as well as rivet expansion had average absolute error of around 2. The damage severity network was also able to perform well with a mean absolute error of about 0.34 for crack length detection and 0.22 for expanded rivet damage. However, the damage localization and severity prediction networks were quite challenging to train in the presence of multiple damages and need further development in the network architecture.

Development of predictive model for biochar surface properties based on biomass attributes and pyrolysis conditions using rough set machine learning

Biochar can be used for environmental remediation, which includes carbon sequestration and soil quality improvement. Biochar is produced from the thermochemical conversion (i.e., pyrolysis) of biomass under inert conditions. However, there are no general rules regarding the relationship between biochar surface properties and biomass physiochemical properties as well as pyrolysis conditions. Machine learning (ML) algorithms can be used to investigate the relation between data sets and deliver useful decision output. In this work, rough set machine learning (RSML) was applied to generate a prediction model of biochar surface properties based on decisional attributes. The prediction model is a rule-based model that contains if-then rules to classify properties by fulfilling conditions. As a result, the specific surface area, pore volume, and pore diameter of biochar were found to be strongly influenced by pyrolysis conditions which includes temperature and retention time as well as biomass attributes including volatile matter, fixed carbon, and ash content. The results generated from RSML showed that the preferred range for pyrolysis temperature to produce biochar with desired surface properties is in between 425 °C and 625 °C, as well as retention time lower than 0.75 h.

An Interpretable Deep Learning Method for Identifying Extreme Events under Faulty Data Interference

Structural health monitoring systems continuously monitor the operational state of structures, generating a large amount of monitoring data during the process. The structural responses of extreme events, such as earthquakes, ship collisions, or typhoons, could be captured and further analyzed. However, it is challenging to identify these extreme events due to the interference of faulty data. Real-world monitoring systems suffer from frequent misidentification and false alarms. Unfortunately, it is difficult to improve the system’s built-in algorithms, especially the deep neural networks, partly because the current neural networks only output results and do not provide an interpretable decision-making basis. In this study, a deep learning-based method with visual interpretability is proposed to identify seismic data under sensor faults interference. The transfer learning technique is employed to learn the features of seismic data and faulty data with efficiency. A post hoc interpretation algorithm, termed Gradient-weighted Class Activation Mapping (Grad-CAM), is embedded into the neural networks to uncover the interest regions that support the output decision. The in situ seismic responses of a cable-stayed long-span bridge are used for method verification. The results show that the proposed method can effectively identify seismic data mixed with various types of faulty data while providing good interpretability.

The networked micro-decision context: a new lens on transformative urban governance

Recent large-scale societal disruptions, from the COVID-19 pandemic to intensifying wildfires and weather events, reveal the importance of transforming governance systems so they can address complex, transboundary, and rapidly evolving crises. Yet current knowledge of the decision-making dynamics that yield transformative governance remains scant. Studies typically focus on the aggregate outputs of government decisions, while overlooking their micro-level underpinnings. This is a key oversight because drivers of policy change, such as learning or competition, are prosecuted by people rather than organizations. We respond to this knowledge gap by introducing a new analytical lens for understanding policymaking, aimed at uncovering how characteristics of decision-makers and the structure of their relationships affect their likelihood of effectuating transformative policy responses. This perspective emphasizes the need for a more dynamic and relational view on urban governance in the context of transformation.

Selecting renewable desalination using uncertain data: an MCDM framework combining mixed objective weighting and interval MARCOS

Abstract Using renewable energy to drive desalination is increasingly favored to augment water supply, given its cost advantage and energy saving. Divergent renewables can be integrated into different desalination technologies, resulting in the selection of an appropriate renewable desalination being a challenge. This work proposes a multi-criteria decision-making framework to evaluate renewable desalination alternatives from the perspective of multi-dimensional consideration and data uncertainty by developing a mixed objective weighting (MOW) method and extending the method of MARCOS (measurement of alternatives and ranking according to a compromise solution) into uncertain conditions. Mathematical contributions can be found in the two methods, i.e. the MOW improves the objectivity and fairness in the weighting result by considering the dispersions and correlations among the criteria's performances, and interval MARCOS guarantees stability in the ranking result while well preserving the uncertain data. An illustrative case concerning four renewable desalination alternatives is used to test the feasibility of the framework. After implementing comparisons regarding the weighting result and ranking sequence, the effectiveness of the involved methods is confirmed. In conclusion, by fully utilizing objective data concerning renewable desalination while eliminating subjective interference, the developed framework can offer a rational and reliable decision output.

Artificial Intelligence Can Guide Antibiotic Choice in Recurrent UTIs and Become an Important Aid to Improve Antimicrobial Stewardship.

A correct approach to recurrent urinary tract infections (rUTIs) is an important pillar of antimicrobial stewardship. We aim to define an Artificial Neural Network (ANN) for predicting the clinical efficacy of the empiric antimicrobial treatment in women with rUTIs. We extracted clinical and microbiological data from 1043 women. We trained an ANN on 725 patients and validated it on 318. The ANN showed a sensitivity of 87.8% and specificity of 97.3% in predicting the clinical efficacy of empirical therapy. The previous use of fluoroquinolones (HR = 4.23; p = 0.008) and cephalosporins (HR = 2.81; p = 0.003) as well as the presence of Escherichia coli with resistance against cotrimoxazole (HR = 3.54; p = 0.001) have been identified as the most important variables affecting the ANN output decision predicting the fluoroquinolones-based therapy failure. A previous isolation of Escherichia coli with resistance against fosfomycin (HR = 2.67; p = 0.001) and amoxicillin-clavulanic acid (HR = 1.94; p = 0.001) seems to be the most influential variable affecting the output decision predicting the cephalosporins- and cotrimoxazole-based therapy failure. The previously mentioned Escherichia coli with resistance against cotrimoxazole (HR = 2.35; p < 0.001) and amoxicillin-clavulanic acid (HR = 3.41; p = 0.007) seems to be the most influential variable affecting the output decision predicting the fosfomycin-based therapy failure. ANNs seem to be an interesting tool to guide the antimicrobial choice in the management of rUTIs at the point of care.

Poison Neural Network-Based mmWave Beam Selection and Detoxification With Machine Unlearning

Deep neural network-based learning methods have been considered promising techniques used in beam selection problems. However, existing research ignores the peculiar vulnerabilities of neural networks. The adversaries can use data poisoning to embed predefined triggers into a model during training time such that the neural network-based beam model may make an incorrect output decision of a test example when patched with the trigger. Data poisoning offers attackers the possibility to build backdoors. The goal of backdoors is often unethical, such as giving users a poor experience by manipulating infected models to output inappropriate beams. In this paper, first, we introduce a simple backdoor attack method by using data poisoning in a mmWave beam selection system. By numerical simulations, we verify that this poisoning attack is effective for neural networks with different structures. In addition, we explore the effect of poisoned data volume on the effect of backdoor attacks. The results show that the backdoor can be successfully implanted into the beam selection neural network. Besides, we fine-tune the trained model for a new wireless communication environment, and the results show that backdoors still exist even when the model is tuned with data from new scenarios. Then, we propose a machine unlearning solution to mitigate the backdoor of the trained beam selection model. The problem of eliminating backdoors is modeled as a minimax optimization problem. We propose a novel adversarial unlearning method along with label smoothing to solve the backdoor removal problem. We compared the proposed backdoor elimination method with the classical fine-tuning elimination method and the neural network pruning method through numerical simulations. The results show that the fine-tuning and the pruning methods cannot effectively remove the backdoor. The proposed machine unlearning method can make the trained model forget about the backdoor under the condition that the performance of the benign task (beam selection tasks when the trigger does not appear) is guaranteed to be slightly degraded. In summary, our work illustrates that data poisoning-based backdoor attacks may exist in wireless networks, and we propose a scheme to eliminate backdoors.