Black-box Design Research Articles

Feature selection technique on convolutional neural network – multilabel classification task

Automated text-based recommendation, an artificial intelligence development, finds application in document analysis like job resumes. The classification of job resumes poses challenges due to the ambiguity in categorizing multiple potential jobs in a single application file, termed multi-label classification, deep learning, particularly convolutional neural networks (CNN), offers flexibility in enhancing feature representations. Despite its robust learning capabilities, the black-box design of deep learning lacks interpretability and demands a substantial number of parameters, requiring significant computational resources. The primary challenge in multilabel learning is the ambiguity of labels not fully explained by traditional equivalence relations. To address this, the research employs feature selection techniques, specifically the Chi-square method. The goal is to reduce features in deep learning models while considering label relevance in multi-label text classification, easing computational workload while preserving model performance. Experimental tests, both with and without the Chi-square feature selection technique on the dataset, underscore its substantial impact on the classification model's ability. The conclusion emphasizes the influence of the Chi-square feature selection technique on performance and computational time. In summary, the research underscores the importance of balancing computational efficiency and model interpretability, especially in complex multi-label classification tasks like job applications.

Deep Unfolded Network with Intrinsic Supervision for Pan-Sharpening

Existing deep pan-sharpening methods lack the learning of complementary information between PAN and MS modalities in the intermediate layers, and exhibit low interpretability due to their black-box designs. To this end, an interpretable deep unfolded network with intrinsic supervision for pan-sharpening is proposed. Building upon the observation degradation process, it formulates the pan-sharpening task as a variational model minimization with spatial consistency prior and spectral projection prior. The former prior requires a joint component decomposition of PAN and MS images to extract intrinsic features. By being supervised in the intermediate layers, it can selectively provide high-frequency information for spatial enhancement. The latter prior constrains the intensity correlation between MS and PAN images derived from physical observations, so as to improve spectral fidelity. To further enhance the transparency of network design, we develop an iterative solution algorithm following the half-quadratic splitting to unfold the deep model. It rigorously adheres to the variational model, significantly enhancing the interpretability behind network design and efficiently alternating the optimization of the network. Extensive experiments demonstrate the advantages of our method compared to state-of-the-arts, showcasing its remarkable generalization capability to real-world scenes. Our code is publicly available at https://github.com/Baixuzx7/DISPNet.

KT-EGO: a knowledge transfer assisted efficient global optimization algorithm for solving high-dimensional expensive black-box problems

Many engineering problems involve optimizing a high-dimensional expensive black-box (HEB) design space. To solve such problems efficiently, a knowledge transfer (KT) assisted efficient global optimization (EGO) algorithm is proposed, called the KT-EGO, which extends the EGO algorithm for solving problems over higher dimensions (i.e.d>20). Specifically, the original design space is divided into several low-dimensional subset design spaces. More importantly, in order to extract information from the subset design spaces to accelerate the progress of full optimization, a surrogate-based data fusion strategy is proposed in the KT-EGO. And further, a searching strategy with an adaptive variable range is devised to enhance the exploitation of promising areas. To show the effectiveness of the proposed algorithm, it is compared against state-of-the-art algorithms over 12 benchmark functions and a 28-dimensional engineering optimization for the design of a compressor blade, which fully validates the effectiveness of the KT-EGO for solving HEB problems.

Research on the Sensory Feeling of Product Design for Electric Toothbrush Based on Kansei Engineering and Back Propagation Neural Network

<p>Over the years, China’s electric toothbrush market has been expanding. Consumers pay more attention to the sensory feeling of product shape, under the premise of product function satisfaction. Therefore, this research collected 215,827 product reviews made by consumers online and 200 samples of varying electric toothbrush samples using a web crawler. Then, 3 groups of representative perceptual words were obtained from the extraction of numerous reviews via Word2vec, factor analysis and hierarchical cluster analysis. Meanwhile, with the help of morphological analysis, design elements of sample shape were de-structured on the 32 representative samples that were extracted from the collected sample using multi-dimensional scaling and hierarchical cluster analysis. On this basis, consumers’ perceptual images were measured using semantic differential scale with 415 valid samples acquired in total. Finally, two relationship models between product design elements and consumers’ perceptual images were established by quantitative theory type I (QTTI) and back propagation neural network. By comparison, the QTTI model has more accurate prediction. This study provides defined design indexes and references for designers’ black box design patterns through establishing an effective model via combining web crawler technology and systematic analysis.</p> <p> </p>

A Nested Weighted Tchebycheff Multi-Objective Bayesian Optimization Approach for Flexibility of Unknown Utopia Estimation in Expensive Black-Box Design Problems

Abstract We propose a nested weighted Tchebycheff Multi-objective Bayesian optimization (WTB MOBO) framework where we built a regression model selection procedure from the ensemble of models, toward better estimation of the uncertain parameters (utopia) of the weighted Tchebycheff expensive black-box multi-objective function. In our previous work, a weighted Tchebycheff MOBO approach has been demonstrated which attempts to estimate the model parameters (utopia) in formulating the acquisition function of the weighted Tchebycheff multi-objective black-box functions, through calibration using an a priori selected regression model. However, the existing MOBO model lacks flexibility in selecting the appropriate regression models given the guided sampled data and, therefore, can under-fit or over-fit as the iterations of the MOBO progress. This ultimately can reduce the overall MOBO performance. As, in general, it is too complex to a priori guarantee a best model, this motivates us to consider a portfolio of different families (simple-to-complex) of predictive models that have been fitted with current training data guided by the WTB MOBO, and the best model is selected following a user-defined prediction root-mean-square error-based approach. The proposed approach is implemented in optimizing a thin tube design under constant loading of temperature and pressure, minimizing the risk of creep-fatigue failure and design cost. Finally, the nested WTB MOBO model performance is compared with different MOBO frameworks with respect to accuracy in parameter estimation, Pareto-optimal solutions, and function evaluation cost. This approach is generalized enough to consider different families of predictive models in the portfolio for best model selection, where the overall design architecture allows for solving any high-dimensional (multiple functions) complex black-box problems and can be extended to any other global criterion multi-objective optimization methods where prior knowledge of utopia is required.

Data Augmentation Empowered Neural Precoding for Multiuser MIMO With MMSE Model

Precoding design exploiting deep learning methods has been widely studied for multiuser multiple-input multiple-output (MU-MIMO) systems. However, conventional neural precoding design applies black-box-based neural networks which are less interpretable. In this letter, we propose a deep learning-based precoding method based on an interpretable design of a neural precoding network, namely iPNet. In particular, the iPNet mimics the classic minimum mean-squared error (MMSE) precoding and approximates the matrix inversion in the design of the neural network architecture. Specifically, the proposed iPNet consists of a model-driven component network, responsible for augmenting the input channel state information (CSI), and a data-driven sub-network, responsible for precoding calculation from this augmented CSI. The latter data-driven module is explicitly interpreted as an unsupervised learner of the MMSE precoder. Simulation results show that by exploiting the augmented CSI, the proposed iPNet achieves noticeable performance gain over existing black-box designs and also exhibits enhanced generalizability against CSI mismatches.

Ayurveda treatment as black box design in the management of Visarpa- Retrospective case study

Visarpa is one of skin disease explained by most of the Acharyas. A 24-years old female came to Panchakarma OPD of Dhanvantari Ayurveda College Hospital and Research Centre, Siddapur, Uttara Kannada. With the complaints of blisters on the right-side of neck, severe burning sensation, itching, severe pricking pain, redness, pus formation for 3 days. She was successfully treated with external applications and internal medications with Tikta (bitter), Kashaya (astrigent) and Sheeta (cool) drugs. This case study demonstrates that acute condition of Visarpa can be managed well in short duration with Ayurvedic medicines.

Multi-level spatial details cross-extraction and injection network for hyperspectral pansharpening.

Hyperspectral (HS) pansharpening, which fuses the HS image with a high spatial resolution panchromatic (PAN) image, provides a good solution to overcome the limitation of HS imaging devices. However, most existing convolutional neural network (CNN)-based methods are hard to understand and lack interpretability due to the black-box design. In this Letter, we propose a multi-level spatial details cross-extraction and injection network (MSCIN) for HS pansharpening, which introduces the mature multi-resolution analysis (MRA) technology to the neural network. Following the general idea of MRA, the proposed MSCIN divides the pansharpening process into details extraction and details injection, in which the missing details and the injection gains are estimated by two specifically designed interpretable sub-networks. Experimental results on two widely used datasets demonstrate the superiority of the proposed method.

The adoption of deep learning interpretability techniques on diabetic retinopathy analysis: a review.

Diabetic retinopathy (DR) is a chronic eye condition that is rapidly growing due to the prevalence of diabetes. There are challenges such as the dearth of ophthalmologists, healthcare resources, and facilities that are unable to provide patients with appropriate eye screening services. As a result, deep learning (DL) has the potential to play a critical role as a powerful automated diagnostic tool in the field of ophthalmology, particularly in the early detection of DR when compared to traditional detection techniques. The DL models are known as black boxes, despite the fact that they are widely adopted. They make no attempt to explain how the model learns representations or why it makes a particular prediction. Due to the black box design architecture, DL methods make it difficult for intended end-users like ophthalmologists to grasp how the models function, preventing model acceptance for clinical usage. Recently, several studies on the interpretability of DL methods used in DR-related tasks such as DR classification and segmentation have been published. The goal of this paper is to provide a detailed overview of interpretability strategies used in DR-related tasks. This paper also includes the authors’ insights and future directions in the field of DR to help the research community overcome research problems.Graphical abstract

Add-on effect of Ayurvedic treatment protocol for diabetic retinopathy: A randomized controlled clinical study.

Diabetic retinopathy (DR), the leading cause of visual disability in diabetics, is a significant complication of diabetes mellitus. Currently available conventional treatments for DR have certain limitations, considering which Ayurvedic treatment protocol was designed. The aim of this study was to evaluate the clinical efficacy of the Ayurvedic treatment protocol for DR. This was a randomized, controlled, black box design clinical study conducted from April 2016 to September 2017 by the department of Shalakya Tantra of a tertiary academic hospital in Western India. A hundred patients of DR in the age group 30-70 years were randomly divided into two groups by simple random sampling using computer-generated random number tables. In the trial group (n = 70), the preparatory phase included Dipana-Pachana (stomachic and digestant), Koshtha Shodhana (mild therapeutic purgation), and Shiro Virechana (eliminative nasal medication). The treatment phase included Marsha Nasya (nasal medication) and Pratimarsha Nasya (nasal medication of mild dose) with Durvadi Ghrita, Takra Dhara (pouring medicated buttermilk over the scalp) with Siddha Takra, and intake of Rasayana Yoga (treatment duration - 3 months). In the control group (n = 30), patients were kept under conservative treatment and observed during the trial period of 3 months. Patients of both groups continued with their treatment for diabetes and DR if any. Two follow-ups were done at an interval of 15 days. The primary outcomes were objective signs like best-corrected visual acuity (BCVA); ophthalmoscopic signs such as superficial hemorrhages, dot-blot hemorrhages, hard exudates, cotton wool spots, neovascularization disc, neovascularization elsewhere, and fibrovascular proliferation; subjective symptoms such as diminished vision, blurred vision, frequent changes in presbyopia glasses, perception of flashes of light, floaters, and problem for dark adaptation. The secondary outcomes were fasting blood sugar (FBS), postprandial blood sugar (PPBS), urine sugar, serum cholesterol, hemoglobin (Hb), glycosylated HbA1C, liver function test, and renal function test outcomes were assessed before and after the treatment. Ninety participants were included in the analysis of the primary outcome (62 in the trial and 28 in the control group). The trial group provided better results which were statistically significant on dot-blot hemorrhages, superficial hemorrhages, hard exudates, BCVA, FBS, and serum cholesterol. Both the groups provided almost similar effects in PPBS, Hb, HbA1C, and urine sugar which were statistically insignificant. Adverse effects were not reported in any of the patient among either groups. Ayurvedic treatment protocol is safe and effective in DR.

Accurate and Lightweight Image Super-Resolution With Model-Guided Deep Unfolding Network

Deep neural networks (DNNs) based methods have achieved great success in single image super-resolution (SISR). However, existing state-of-the-art SISR techniques are designed like black boxes lacking transparency and interpretability. Moreover, the improvement in visual quality is often at the price of increased model complexity due to black-box design. In this paper, we present and advocate an explainable approach toward SISR named model-guided deep unfolding network (MoG-DUN). Targeting at breaking the coherence barrier, we opt to work with a well-established image prior named nonlocal auto-regressive model and use it to guide our DNN design. By integrating deep denoising and nonlocal regularization as trainable modules within a deep learning framework, we can unfold the iterative process of model-based SISR into a multi-stage concatenation of building blocks with three interconnected modules (denoising, nonlocal-AR, and reconstruction). The design of all three modules leverages the latest advances including dense/skip connections as well as fast nonlocal implementation. In addition to explainability, MoG-DUN is accurate (producing fewer aliasing artifacts), computationally efficient (with reduced model parameters), and versatile (capable of handling multiple degradations). The superiority of the proposed MoG-DUN method to existing state-of-the-art image SR methods including RCAN, SRMDNF, and SRFBN is substantiated by extensive experiments on several popular datasets and various degradation scenarios.

Optimization of expensive black-box problems via Gradient-enhanced Kriging

This paper explores the use of Gradient-enhanced Kriging for optimization of expensive black-box design problems, which is not completely limited by the conventional Efficient Global Optimization algorithm framework. Specifically, we give the best linear unbiased predictor and mean squared prediction error of the partial derivatives of Gradient-enhanced Kriging and then propose a measure named “Approximate Probability of Stationary Point” to estimate the approximate probability of a candidate infill point be a stationary point of the underlying function. When it comes to the selection of infill point, we not only maximize the well-known Expected Improvement but also evaluate the Approximate Probability of Stationary Point as a “double-check” step. Then the infill decision is made according to the extent of consistency between these two quantities. Furthermore, to examine whether the optimization process will gain from sparing more costs for response evaluation, we investigate also the cases that the gradient evaluation step is conditionally skipped in some iterations. Three new infill criteria are proposed and experimented with three analytical test functions and an airfoil optimal shape design. Results show that the optimization performance can be improved by exploiting the auxiliary gradient information in the proposed way.

Modifying games with ChoiCo: Integrated affordances and engineered bugs for computational thinking

AbstractAlthough there is wide rhetoric that programming should be learnt by all as an element of computational thinking (CT), in practice, it is mostly implemented narrowly as an end in itself consisting of routine practice and traditional testing of the ability to code. This paper discusses a way in which programming could be seen through a wider integrated pedagogical approach as jointly cultivating meaning making of computational concepts in conjunction with the adoption of practices and strategies in a relevant meaningful context for learners. We elaborate on a case of learning to code through digital game modding where programming and other computational concepts coexist. Our design frame includes the principles of black and white box designs, of students as de‐buggers of engineered half‐baked games and of games embedding both concepts and values in simulations involving wider contested complex issues. We use our analysis of the meaning making of students as they debug a choice‐driven simulation game specially designed to bring this integrated kind of learning to the fore. We show elements of context‐aware integrated CT connecting otherwise fragmented areas such as databases, block‐based programming, Geographical Information Systems design. Practitioner NotesWhat is already known about this topic Computational thinking is considered an essential element of digital citizenship. In this context, it has been portrayed as an educationally strategic 21st century competency that all students should acquire. According to several studies it involves the development of both concepts and practices that are necessary for computational problem solving. Concepts such as variables, conditionals, data structures and practices such as decomposition, abstraction, generalisation. However, most of the pedagogical approaches for CT focus rather narrowly on computational concepts and engage students with closed quizzes and puzzles, leaving aside practices. Not surprisingly, students have difficulties in describing and using computational practices (eg, abstraction, decomposition, pattern recognition) even if they have sufficient knowledge on programming concepts. Computational design activities, eg, game or simulation design, require high‐order skills and knowledge which are complex and inaccessible by the young and inexperienced students. There is need for approaches that would foster students to express and develop meanings on both practices and concepts through an integrated context. What this paper adds It suggests the design of tools with affordances aiming at students' integrated engagement with concepts and practices. It provides an example of a tool integrating affordances connected to diverse computational concepts such as those pertaining to programming, data handling and Geographical Information Systems (GIS). It discusses design principles connected to the embedding of integrating affordances, such as making use of the context of designing and playing digital games. It refers to the embedding of domain concepts in realistic contexts of wider complex real life issues such as sustainability or dietary habits. It provides in‐depth analysis of student engagement with such tools and uses this data to elaborate the arguments and issues related to this design approach. Implications for practice and/or policy There is value in disconnecting CT from positivist diagnostic approaches related to respective concepts. There is need for more resources into the design and development of digital media embedding affordances for concepts and practices while maintaining relevance and interest for their users. More examples are needed and analyses of student learning processes and how they may affect their digital citizenship more widely. These approaches need to combine individualistic and collective approaches to human learning.

Negotiation game for joint IT and energy management in green datacenters

As the power demand of datacenters is increasing sharply, a promising solution is to power datacenters locally by renewable energies. However, one of the main challenges when operating such green datacenters is to conciliate the intermittent power supply and the power demand. To deal with this problem, we view the green datacenter as two sub-systems, namely, Information Technology (IT) sub-system which consumes energy, and electrical sub-system which supplies energy. The objective is to find an efficient trade-off between the power demand and power supply, respecting the operational requirements of both sub-systems (i.e., the requirements on utility, or monetary gain, which includes monetary revenue and monetary cost). First, we analyze the problem by a black-box approach. In this approach, the models of the two sub-systems are unknown to each other, and the two sub-systems negotiate by exchanging their power preferences. However, we found that the black-box approach cannot guarantee stable solutions in term of execution time and generational distance (which is the distance between a solution and the Pareto front). Then we introduce a semi black-box approach, in which the two sub-systems are modeled as the buyer and the supplier in a buyer–supplier negotiation game. We propose an algorithm that allows the buyer and supplier to negotiate, seeking for an efficient trade-off between the power demand and power supply. The analytical results show that the semi black-box algorithm converges to equilibrium, and these results are then confirmed by experimental results. We conduct the experiments by implementing a middleware of a datacenter powered by renewable energies. The experimental results show that the semi black-box algorithm improves significantly the stability, quality of service (QoS) and utility of the datacenter, compared to other algorithms. In term of stability, compared to the black-box algorithm, the semi black-box algorithm reduces the standard deviation of execution time and generational distance by 23 and 27 times, respectively. In term of QoS and utility, the semi black-box algorithm outperforms the algorithms that do not consider joint IT-energy management, as well as the algorithms that do not utilize a semi black-box design.

PRESSURE CONTROLS: DATA FORCES CREEPING THROUGH FINGERTIP COMMANDS

AI methods and ubiquitous data sensors have enabled a new algorithmic quantification of affect with the possibility to detect or verify users’ identities, characteristics, emotional states, and physical traits. By scrutinizing how transient datasets are produced by user applied pressure on touch-screens (via fingertip commands) this paper showcases how sensory technology creeps into users’ everyday life with potential implementations connected to a series of emerging data issues engineered by a black-box design: one which obfuscates data production and precludes user consent under the disguise of “non-intrusive” features. Thereby, this paper explores the limits of user-based interrogation of black-boxes by researching tactile modes of operation, as a subset of behavioural biometrics, and sensors that register force in touch analysis and haptic technologies.
 Presenting a citation analysis of biometric techniques around the proposed usage of pressure; the authors offer a case-study examination of zinc-based force-sensing materials that are cost-effective and scalable to ubiquitous-computing and a prototype developed using ‘each pixel as a sensor’. By combining these approaches, this paper argues that such developments constitute a phenomenological shift away from users’ perception to data infrastructures working as assemblages of hidden technical sensations, and there is a need to expose these complex networks to afford some grasp, if not direct agency, over their micro temporal operation. This work aims not simply to theorise, but to help reveal ways users may revise, embrace, resist, subvert or even live data practices that operate unlike conventional data harvesting techniques.

Energy Management System in Battery Electric Vehicle Based on Fuzzy Logic Control to Optimize the Energy Consumption in HVAC System

This particular research has been conducted at University Putra Malaysia (UPM) to investigate the applicability of fuzzy logic technique in enhancing the energy management of battery-powered electric vehicle (BEV). Because of the increasing demand of BEVs, there is a need to increase the battery power while fulfilling the conflicting of battery power needs and power consuming needs for motor and auxiliaries (such as HVAC). The balance between keeping the comfort of HVAC use and increasing the battery range is complicated and the support of Artificial Intelligence can be useful. The study integrated an energy management system, which is using the designed Fuzzy Logic strategy, to enhance the drain of battery power capacity. The simple black box design of the EMS system has two inputs, State of Charge (SoC) and Speed, and three outputs, Heated Seats, Front HVAC, and Rear HVAC. The membership functions of output fuzzy of the Front HVAC, which shows that every output has three equal categories, 1/3, 2/3, 3/3, associated with low, mid, and high. The same approach of membership functions is applied for Rear HVAC, and Heated Seat. The three outputs of the HVAC system assumed to have a constant load of 1000 Watt each and have three equal categories, low, medium, and high. The fuzzy logic design uses a strategy of nine rules. The simulation is applied on MATLAB Simulink environment and the tests are using two driving cycles, the New European Driving Cycle (NEDC) and Japan 10-15. The results show that using the managed HVAC strategy can increase the battery driving range by 9.8-20.4% compared with the full-unmanaged HVAC strategy.

The Green Architecture Approach to Natural School at Suburb

Natural school is an educational institution that can answer the environmental problems and rigid education system problems with natural nuances. The green architecture approach is used as a design approach to achieve natural school goals that provide an understanding of how awareness of nature is applied in the school. The green architectural approach can be applied by the use of environmentally friendly local materials, natural ventilation, environmentally friendly waste management, and the use of the minimum natural resource. This study uses a glass box and a black box design method. The result of this study is a natural school design that is environmentally friendly and healthy in the suburb of Medan, Indonesia.

Selecting evolutionary algorithms for black box design optimization problems

An algorithm selection method for black box design optimization problems is reported. It uses a simple and natural principle to select an algorithm set from a set of algorithm candidates. A set of benchmark problems is given, and the performance of each algorithm in the set is recorded in a knowledge base. Given an unknown problem, the default algorithm is run. An algorithm–problem feature is proposed and used to map to the most similar benchmark problem. Then the best algorithm for solving the problem is used in the second run. This process iterates until n runs have been made. The best result out of n runs is returned as the solution. Experimental results reveal that the algorithm–problem feature is a good problem identifier. Results are also reported when (1) both the training and testing set are the set of benchmark problems; and (2) the training set is the set of benchmark problems but the testing set is a set of real-world benchmark problems. The method works well on both scenarios. It attains almost the same performance as the best algorithm, and has better performance compared with random selection. As the best algorithm cannot be known a priori, the results confirm that the algorithm selection mechanism is effective. The performance of the algorithm as a function of n and the case when n is equal to two is also studied.

An Innovative Approach to Mechanical Product Designs - A Case Study on Computer Cooling Fan Design

With advances in various technologies, it is critical to create a product styling design that can trigger consumers’ willingness to buy. Various studies have been emphasizing on product styling designs that can grasp a consumer’s mind. Average people at this information age have higher demands toward product designs. When considering a purchase, people’s first consideration is whether a product is of high quality at a lower price. Product competitions on the market are affected interactively by globalization and localization and the awareness of humanization and personalization. The conventional black box design approach makes it difficult for a designer to meet consumer demands effectively. Via the framework of mechanical product innovation design approach in this study, patent specifications were reviewed and analyzed in order to decompose the essential elements and their corresponding functions. Correlation tables were built and product structural formula could be determined. More functions were included by replacing elements in order to achieve the completeness of the structural formula. For demonstrating the benefits of this approach, the design of computer cooling fan housing was chosen as a case study. The results indicated that the proposed design approach can effectively assist in the development of product innovative designs.

Efficacy of Multi Unani regimen in functional independence recovery of post stroke patients and #8211; A black-box design

Efficacy of Multi Unani regimen in functional independence recovery of post stroke patients and #8211; A black-box design