White-box Model Research Articles

Skin and Proximity Effect Calculation of a System of Rectangular Conductors Using the Proper Generalized Decomposition Technique

This paper presents the application of a numerical approach known as proper generalized decomposition (PGD) to calculate the per-unit length (PUL) ac resistance of rectangular conductors. PGD has been successfully used in areas such as fluid mechanics and biomedical applications. It solves a partial differential equation (PDE) by decomposing the answer into a set of unknown one-dimensional (1D) functions in an iterative approach until it reaches a predetermined convergence. In this paper, a frequency-dependent meshing scheme is employed in the PGD technique at each frequency to properly take skin and proximity effects into account. One of the main advantages of PGD over traditional numerical approaches such as finite element or finite difference methods is that it confines the answers within a set of one-dimensional functions, which require fewer computational resources. Different examples of single and multiple rectangular conductors are considered to study skin and proximity effects. The PGD results are compared with those obtained using a commercial finite element method (FEM) software to verify the accuracy of the model. This approach can be used in applications such as white box modeling of transformers, EMC analysis, hairpin winding design used in electric vehicles, and busbar simulation.

Enhancing decanter centrifuge process design with data‐driven material parameters in multi‐compartment modeling

AbstractPredicting the separation performance of decanter centrifuges is challenging due to dynamic events within the apparatus. Current methods for designing decanter centrifuges rely on simplified models, often leading to inaccuracies. Consequently, manufacturers must perform time‐intensive pilot scale experiments to derive their own correction factors. Growing computing power sparks interest in alternative modeling strategies. Grey box models (GBM) combine mechanistic white box models (WBM) and data‐driven black box models (BBM), with the optimal structure (parallel or serial) varying by application. For modeling decanter centrifuges, we propose a serial GBM that comprises an artificial neural network that outputs unknown material parameters into a first‐principle multi‐compartment model. Comparing this approach to alternative data‐driven modeling strategies (pure BBM, parallel GBM), we conclude that the serial GBM excels in terms of extrapolation, prediction ability, and transparency while also enabling a better comprehension of the separation process.

Turbine-Level Possible Power Prediction using Farm-Wide Spatial Information through Similarity-Based Multivariate Gaussians

Wind farms usually comprise several turbines of the same type in proximity to one another. Therefore, similarities exist between the power production of specific turbines within the wind farm over time. Considering this, it is possible to find a way to express the similarity between turbines and exploit their properties to find a formulation of the expected behavior of a turbine. With this estimation of possible power output, one can analyze the losses generated by curtailments or transients with a higher precision. Based on this, a probabilistic model is proposed that can be used to calculate energy losses due to maintenance or environmental reasons. On top of that, heavy deviations in the behavior of specific wind turbines can be detected on high-frequency (1-second) data. Overall, the goal of this work is to predict possible power on high-frequency SCADA data using a statistical white-box modeling approach. The presented method is based on a probabilistic framework, constructing a system of linear combinations that permits analytically tracking the expected behavior of a turbine, even if it is non-operational for a specific amount of time. The methodology includes two parts, the first one is the use of data-driven power curves, and the second one consists of an inferential framework based on the environmental conditions of the farm. Results show that the method presented performs better than the manufacturer’s power curves under specific wind speeds and wind directions.

Explainable hypoglycemia prediction models through dynamic structured grammatical evolution

Effective blood glucose management is crucial for people with diabetes to avoid acute complications. Predicting extreme values accurately and in a timely manner is of vital importance to them. People with diabetes are particularly concerned about suffering a hypoglycemia (low value) event and, moreover, that the event will be prolonged in time. It is crucial to predict hyperglycemia (high value) and hypoglycemia events that may cause health damages in the short term and potential permanent damages in the long term. This paper describes our research on predicting hypoglycemia events at 30, 60, 90, and 120 minutes using machine learning methods. We propose using structured Grammatical Evolution and dynamic structured Grammatical Evolution to produce interpretable mathematical expressions that predict a hypoglycemia event. Our proposal generates white-box models induced by a grammar based on if-then-else conditions using blood glucose, heart rate, number of steps, and burned calories as the inputs for the machine learning technique. We apply these techniques to create three types of models: individualized, cluster, and population-based. They all are then compared with the predictions of eleven machine learning techniques. We apply these techniques to a dataset of 24 real patients of the Hospital Universitario Principe de Asturias, Madrid, Spain. The resulting models, presented as if-then-else statements that incorporate numeric, relational, and logical operations between variables and constants, are inherently interpretable. The True Positive Rate and True Negative Rate metrics are above 0.90 for 30-minute predictions, 0.80 for 60 min, and 0.70 for 90 min and 120 min for the three types of models. Individualized models exhibit the best metrics, while cluster and population-based models perform similarly. Structured and dynamic structured grammatical evolution techniques perform similarly for all forecasting horizons. Regarding the comparison of different machine learning techniques, on the shorter forecasting horizons, our proposals have a high probability of winning, a probability that diminishes on the longer time horizons. Structured grammatical evolution provides advanced forecasting models that facilitate model explanation, modification, and retesting, offering flexibility for refining solutions post-creation and a deeper understanding of blood glucose behavior. These models have been integrated into the glUCModel application, designed to serve people with diabetes.

Transferring Black-Box Decision Making to a White-Box Model

In the rapidly evolving realm of artificial intelligence (AI), black-box algorithms have exhibited outstanding performance. However, their opaque nature poses challenges in fields like medicine, where the clarity of the decision-making processes is crucial for ensuring trust. Addressing this need, the study aimed to augment these algorithms with explainable AI (XAI) features to enhance transparency. A novel approach was employed, contrasting the decision-making patterns of black-box and white-box models. Where discrepancies were noted, training data were refined to align a white-box model’s decisions closer to its black-box counterpart. Testing this methodology on three distinct medical datasets revealed consistent correlations between the adapted white-box models and their black-box analogs. Notably, integrating this strategy with established methods like local interpretable model-agnostic explanations (LIMEs) and SHapley Additive exPlanations (SHAPs) further enhanced transparency, underscoring the potential value of decision trees as a favored white-box algorithm in medicine due to its inherent explanatory capabilities. The findings highlight a promising path for the integration of the performance of black-box algorithms with the necessity for transparency in critical decision-making domains.

Conceptualization of Intelligent Control System for Humanitarian Demining Robotic Complexes Based on Verbal Methods

Introduction. Humanitarian demining is characterized by growing attention to the problems of creating and using robotic systems. The most important aspect of their use is the control method.Problem Statement. At present, the fi rst generation robotic complexes (controlled devices) have been the mostcommon, the second generation complexes (semi-autonomous devices) have been improving. To switch to the third generation complexes (autonomous devices), it is necessary to develop intelligent control systems based on artificial intelligence technologies. The most used for system development are quantitative methods, but such models are “black box models” that do not provide complete clarity about their behavior.Purpose. To develop conceptual model of intelligent control systems for humanitarian demining robotic complexes based on verbal methods.Materials and Methods. Formal logic methods, verbal analysis qualitative methods, and BPMN. Results. Conceptual model of intelligent control systems for humanitarian demining robotic complexes and sym bolic models of representing relevant declarative and procedural knowledge based on verbal methods have been developed.Conclusions. The conceptual model makes it possible to formulate Symbolic Models in the notations of selected verbal methods. At the decision-making level, the ordinary classifi cation method by denotation of the terms used by experts in the selected subject area has been chosen. At the executive level, the BPMN method has been chosen to create process diagrams in graphical notation. The verbal methods make it possible to create “white-box models” that unambiguously interpret the dependence of output and input variables and to explain the model behavior. Symbolic models in the notations of selected verbal methods allow the implementation of an intelligent control systems for a specifi c humanitarian demining robotic complex.

CFOA: Exploring transferable adversarial examples by content feature optimization with attention guidance

Deep neural networks are known to be highly susceptible to adversarial examples, created by adding small and carefully-designed perturbations on original images. In black-box scenarios, attacking neural networks is challenging since the information about models is not available. Generating transferable adversarial examples on surrogate models is a viable way. However, current mainstream transfer-based attacks seldomly consider the relationship between transferability and content features of the image. To effectively enhance the attack success rate at the feature level, we newly propose an attack method called Content Feature Optimization Attack (CFOA). This novel CFOA approach operates on white-box models, aiming to generate transferable adversarial examples by targeting the content feature representations of images in the feature space. The method involves the extraction and parameterization of content features, followed by the iterative generation of feature-level perturbations guided by attention loss and adversarial loss. Ultimately, the adversarial examples generated by CFOA exhibit strong transferability on other black-box models. Experimental results show that our method produces adversarial examples with higher success rates in transferability compared to state-of-the-art methods, and importantly, they can effectively evade certain defense mechanisms, presenting a challenge for adversarial defense.

Individual-specific postural discomfort prediction using decision tree models

The objective of the current study was to explore the utilization of the decision tree (DT) algorithm to model posture-discomfort relationships at the individual level. The DT algorithm has the advantage that it makes no assumptions about the distribution of data, is robust in representing non-linear data with noise, and produces white-box models that are interpretable. Individual-level modelling is essential for examining individual-specific postural discomfort perception processes and understanding the inter-individual variability. It also has practical applications, including the development of individual-specific digital human models and more precise and informative population accommodation analysis. Individual-specific DT models were generated using postural discomfort rating data for various seated upper body postures to predict discomfort based on postural and task variables. The individual-specific DT models accurately predicted postural discomfort and revealed large inter-individual variability in the modelling results. DT modelling is expected to greatly facilitate investigating the human discomfort perception process.

Advancing Educational Insights: Explainable AI Models for Informed Decision-Making

Abstract: Over the past two decades, the integration of machine learning (ML) techniques within educational frameworks has garnered significant attention. However, despite its widespread adoption, there remains a dearth of research focusing on developing AI systems with a core emphasis on interpretability and explainability. This paper seeks to bridge this gap by proposing an advanced framework that not only predicts students' performance accurately but also offers reliable and interpretable results tailored for career counseling. The framework merges the concepts of ML and Explainable AI (XAI) to address the complexities of career counseling in educational settings. Drawing inspiration from educational data mining, the framework aims to provide insights conducive to students' career growth and decision-making processes. By incorporating MLbased White and Black Box models, the approach analyzes a comprehensive educational dataset comprising academic and employability attributes crucial for job placements and skill development. To enhance interpretability, the framework leverages the NGBoost algorithm, known for its efficiency in prediction modeling. Additionally, it integrates Local Interpretable Modelagnostic Explanations (LIME) and Shapley Additive Explanations (SHAP) methods for providing both local and global explanations of model predictions, ensuring transparency and comprehensibility. Through a series of use cases, the paper showcases the applicability and effectiveness of the framework in providing actionable insights to educators and students alike. In conclusion, this research contributes to the advancement of career counseling in educational contexts by offering a robust and interpretable ML-based framework. By providing transparent insights into students' academic performance and career prospects, the approach facilitates informed decision-making and supports personalized guidance for optimal career trajectories.

SAME: Sample Reconstruction against Model Extraction Attacks

While deep learning models have shown significant performance across various domains, their deployment needs extensive resources and advanced computing infrastructure. As a solution, Machine Learning as a Service (MLaaS) has emerged, lowering the barriers for users to release or productize their deep learning models. However, previous studies have highlighted potential privacy and security concerns associated with MLaaS, and one primary threat is model extraction attacks. To address this, there are many defense solutions but they suffer from unrealistic assumptions and generalization issues, making them less practical for reliable protection. Driven by these limitations, we introduce a novel defense mechanism, SAME, based on the concept of sample reconstruction. This strategy imposes minimal prerequisites on the defender's capabilities, eliminating the need for auxiliary Out-of-Distribution (OOD) datasets, user query history, white-box model access, and additional intervention during model training. It is compatible with existing active defense methods. Our extensive experiments corroborate the superior efficacy of SAME over state-of-the-art solutions. Our code is available at https://github.com/xythink/SAME.

Research on design requirements for passive residual heat removal system of lead cooled fast reactor via model-based system engineering

Traditional text-based system engineering, which has been used in the design and application of passive residual heat removal system (PRHRS) for lead-cooled fast reactors, is prone to several problems such as low development efficiency, long iteration cycles, and model ambiguity. This study aims to effectively address the above-mentioned problems by adopting a model-based system engineering (MBSE) method, which has been preliminarily applied to meet the design requirements of a PRHRS. The design process has been implemented based on the preliminary design of the system architecture and consists of three stages: top-level requirement analysis, functional requirements analysis, and design requirements synthesis. The results of the top-level requirements analysis and the corresponding use case diagram can determine the requirements, top-level use cases, and scenario flow of the system. During the functional requirements analysis, the sequence, activity, and state machine diagrams are used to develop the system function model and provide early confirmation. By comparing these sequence diagrams, the requirements for omissions and inconsistencies can be effectively checked. In the design requirements synthesis stage, the Analytic Hierarchy Process is used to conduct preliminary trade-off calculations on the system architecture, after which a white box model is established during the system architecture design. Through these two steps, the analysis and design of the system architecture are ultimately achieved. The resulting system architecture ensures the consistency of the design requirements. Ultimately, a functional hazard analysis was conducted for a specific incident to validate case requirements and further refine the system architecture. Future research can further reduce the design risk, improve the design efficiency, and provide a practical reference for the design and optimization of PRHRS in digital lead-cooled fast reactors.

A Grey-Box Model of a DC/DC Boost Converter for PV Energy Systems

This paper presents a grey-box model of a DC/DC boost converter for PV energy systems. The proposed model contains a white-box model part and a black-box model part together to prepare a better model for the PV boost converter. The white-box model part is used for knowledge of the circuit by mathematical equations since the black-box model part is used for unknown parameters such as temperature and electromagnetic interference. The black-box part of the proposed model is created by a nonlinear system identification of a real boost converter circuit with an artificial neural network. The precision of the mathematical model and the advantages of the fast prediction ability of the artificial neural network were used together. The proposed grey-box model is compared with the existing state-space and black-box models and experimental results. The results of the study showed that the average correlation between the proposed grey-box model output and the experimental results is 97.52%. Therefore, the proposed model can be used for analyzing DC/DC boost converter output characteristics before field applications.

Improving adversarial transferability through frequency enhanced momentum

The emergence of adversarial examples seriously affects the practical security deployment of convolutional neural networks. The existing attack algorithms perform brilliantly under white-box scenarios, but they show weak transferability when faced with unknown black-box models. Recent studies have revealed that models have different interests in different frequency components of images, and the low-frequency characteristics play a non-negligible role in the decision-making of models. In this article, we present the frequency enhanced momentum iterative attack, called FE-MI-FGSM. Specifically, we use multiple convolution kernels for Gaussian filtering of the image before each gradient update to push the processed image closer to the common decision boundaries of multiple models. Then, the average gradient of the white-box model to these processed images is obtained and used as the perturbation direction to generate adversarial examples with a high success rate of white-box attack and high transferability. The empirical results show that compared with the current mainstream gradient-based methods, our method performs better on both normally trained and adversarially trained models. Besides, our method can combine with gradient-based methods which integrate convergence algorithms or input transformations for the sake of reaching satisfactory improvement of the transferability.

Comparative Study-Based Data-Driven Models for Lithium-Ion Battery State-of-Charge Estimation

Batteries have been considered a key element in several applications, ranging from grid-scale storage systems through electric vehicles to daily-use small-scale electronic devices. However, excessive charging and discharging will impair their capabilities and could cause their applications to fail catastrophically. Among several diagnostic indices, state-of-charge estimation is essential for evaluating a battery’s capabilities. Various approaches have been introduced to reach this target, including white, gray, and black box or data-driven battery models. The main objective of this work is to provide an extensive comparison of currently highly utilized machine learning-based estimation techniques. The paper thoroughly investigates these models’ architectures, computational burdens, advantages, drawbacks, and robustness validation. The evaluation’s main criteria were based on measurements recorded under various operating conditions at the Energy Systems Research Laboratory (ESRL) at FIU for the eFlex 52.8 V/5.4 kWh lithium iron phosphate battery pack. The primary outcome of this research is that, while the random forest regression (RFR) model emerges as the most effective tool for SoC estimation in lithium-ion batteries, there is potential to enhance the performance of simpler models through strategic adjustments and optimizations. Additionally, the choice of model ultimately depends on the specific requirements of the task at hand, balancing the need for accuracy with the complexity and computational resources available and how it can be merged with other SoC estimation approaches to achieve high precision.

86P Artificial intelligence tools to predict early progression of non-small cell lung cancer patients treated with immunotherapy: Preliminary evidence from a single-centre experience of black-box and white-box models pairing

86P Artificial intelligence tools to predict early progression of non-small cell lung cancer patients treated with immunotherapy: Preliminary evidence from a single-centre experience of black-box and white-box models pairing

Low-order gray-box modeling of heating buildings and the progressive dimension reduction identification of uncertain model parameters

Heating buildings have a huge heat capacity and have abilities to accept fluctuating heat supply from renewables. Creating a low-order scalable model that describes the thermal dynamics of coupled heat transfer, charge, and discharge processes among heating buildings is essential to unlocking the thermal storage potential of building mass and achieving space heating decarbonization. White-box models tend to be too complex and black-box models lack scalability. Thus, a low-order gray-box model with uncertain lumped parameters is proposed based on comprehensive heat transfer analyses and necessary assumptions. Thermal mass in heating buildings is divided into four independent heat storage units. The building envelope that can be heated directly or not is regarded as a different heat storage unit. To reduce the search space dimension of model identification, the decision variables are downsized to essential model parameters based on the sensitivity analysis. Moreover, a progressive dimension reduction method is introduced to further improve the probability that the optimization algorithm converges to a global minimum. Case studies show that the calibrated model with global optimal solutions of uncertain model parameters has a good performance on the prediction of the average temperature of indoor air, the average absolute error is only 0.1 °C.

Trusting AI made decisions in healthcare by making them explainable.

In solving the trust issues surrounding machine learning algorithms whose reasoning cannot be understood, advancements can be made toward the integration of machine learning algorithms into mHealth applications. The aim of this paper is to provide a transparency layer to black-box machine learning algorithms and empower mHealth applications to maximize their efficiency. Using a machine learning testing framework, we present the process of knowledge transfer between a white-box model and a black-box model and the evaluation process to validate the success of the knowledge transfer. The presentation layer of the final output of the base white-box model and the knowledge-infused white-box model shows clear differences in reasoning. The correlation between the base black-box model and the new knowledge-infused model is very high, indicating that the knowledge transfer was successful. There is a clear need for transparency in digital health and health in general. Adding solutions to the toolbox of explainable artificial intelligence is one way to gradually decrease the obscurity of black-box models.

Modeling hydrogen solubility in water: Comparison of adaptive boosting support vector regression, gene expression programming, and cubic equations of state

Predicting the solubility of hydrogen (H2) in aqueous solutions is crucial for studying reactions of hydrogen in the formation, which also affects the security and optimal design of hydrogen storage. In this research, five robust machine learning (ML) algorithms, namely adaptive boosting decision tree (AdaBoost-DT), adaptive boosting support vector regression (AdaBoost-SVR), gradient boosting decision tree (GB-DT), gradient boosting support vector regression (GB-SVR), and k-nearest neighbors (KNN) and three powerful white-box techniques, namely gene expression programming (GEP), genetic programming (GP), and group method of data handling (GMDH) were developed to accurately predict H2 solubility in pure and saline water systems. To this aim, a widespread databank containing 427 experimental data points was collected, and temperature, pressure, and salt concentration (mSalt) were considered as input variables. The validity and precision of the developed models were assessed utilizing several statistical and graphical tests. Results demonstrate that the AdaBoost-SVR smart model could obtain a superior performance and provides precise predictions with root mean square error (RMSE) of 0.000115 and determination coefficient (R2) of 0.9973. Among the white-box models, the GEP provided the best results with an RMSE of 0.000362 and an R2 of 0.9542. Although the accuracy of GEP is slightly lower than that of AdaBoost-SVR, it offers explicit and simple mathematical formula for calculating H2 solubility, which is the main advantage of white box models. The results also demonstrated that AdaBoost-SVR outperforms cubic equations of state (EOSs) such as Peng-Robinson (PR), Redlich-Kwong (RK), Soave-Redlich-Kwong (SRK), and Zudkevitch-Joffe (ZJ). Besides, trend analysis showed that AdaBoost-SVR model could match actual trends of H2 solubility change versus temperature and pressure. Finally, outlier detection analysis using the Leverage technique indicated that the majority of data points used for modeling (nearly 94 %) are reliable and placed in the valid zone.

An integrated grey-box model for accurate ship engine performance prediction under varying speed and environmental conditions

The development of ship engine fault diagnosis has led to an increased interest in predicting marine engine performance under various environmental and operating conditions. However, predicting engine performance using a single model is limited due to the various characteristics of ship engines depending on speed and environmental conditions. To address this issue, we propose a grey box model (GBM)-based ship engine performance prediction framework that combines the white box model and black box model (BBM) appropriately for both low- and high-speed operating conditions. The application of data preprocessing techniques, such as clustering and cleaning, under specific engine and environmental conditions, combined with dimensionality reduction techniques (partial least square and principal component) and the use of WBM/BBM models under classified speed conditions, makes the proposed framework accessible to real-world operators through data-driven approaches and domain knowledge. We demonstrate that the proposed framework improves the robustness, accuracy, and efficiency of engine performance predictions by considering the characteristics of each speed in real-world navigation data, compared to using single models.

Improving Transferability of Universal Adversarial Perturbation with Feature Disruption.

Deep neural networks (DNNs) are shown to be vulnerable to universal adversarial perturbations (UAP), a single quasi-imperceptible perturbation that deceives the DNNs on most input images. The current UAP methods can be divided into data-dependent and data-independent methods. The former exhibits weak transferability in black-box models due to overly relying on model-specific features. The latter shows inferior attack performance in white-box models as it fails to exploit the model's response information to benign images. To address the above issues, this paper proposes a novel universal adversarial attack to generate UAP with strong transferability by disrupting the model-agnostic features (e.g., edges or simple texture), which are invariant to the models. Specifically, we first devise an objective function to weaken the significant channel-wise features and strengthen the less significant channel-wise features, which are partitioned by the designed strategy. Furthermore, the proposed objective function eliminates the dependency on labeled samples, allowing us to utilize out-of-distribution (OOD) data to train UAP. To enhance the attack performance with limited training samples, we exploit the average gradient of the mini-batch input to update the UAP iteratively, which encourages the UAP to capture the local information inside the mini-batch input. In addition, we introduce the momentum term to accumulate the gradient information at each iterative step for the purpose of perceiving the global information over the training set. Finally, extensive experimental results demonstrate that the proposed methods outperform the existing UAP approaches. Additionally, we exhaustively investigate the transferability of the UAP across models, datasets, and tasks.