Multiple Domains Research Articles

The convergence of Digital Twins and Distributed Ledger Technologies: A systematic literature review and an architectural proposal

In recent years, the emerging Digital Twin (DT) technology is playing a key role in fostering the transition towards the Industry 4.0. DTs, representing virtual replicas of physical objects, products or processes established thanks to a bidirectional continuous flow of information between the physical and the virtual world, are currently adopted in multiple domains such as manufacturing, aerospace, automotive, energy, construction, smart cities and smart mobility, etc.. DTs live together with the physical system they replicate, receiving the same data and often triggering specific control actions that directly impact on the real system status. When dealing with DTs of complex Cyber-Physical Systems (CPSs), the data sources may be heterogeneous and untrustworthy, which requires the DT to be able to address security issues related to data in transit from/to the physical twin and data at rest. A possible way to address these data security issues consists in adopting Distributed Ledger Technologies (DLTs) which provide several security guarantees on data through cryptographic hashing techniques. In this work, we present a three-fold contribution: (i) we discuss the results of a Systematic Literature Review (SLR) on the state-of-the-art of the research related to the integration between DLT and Digital Twins, aimed at clarifying relevant aspects such as, among others, what is the favourite DLT choice in existing proposals or what is the type of DT-related information to store on-chain; (ii) leveraging the SLR results and other related research, we propose an architectural framework for the integration of DT and DLTs (more specifically, blockchains); (iii) we validate the proposed architecture by means of two proof-of-concept implementations leveraging different technological stacks and taking into account two different operational scenarios. Finally, we conduct a requirements coverage analysis and compare the PoCs through a coverage matrix.

A Porphyrinic Metal-Organic Framework with Cooperative Adsorption Domains for PFAS Removal from Water.

The removal of toxic poly- and perfluoroalkyl substances (PFAS) as persistent pollutants from wastewater is imperative but challenging for water remediation. Many adsorbents including activated carbon, biochar, and clay minerals have been investigated for PFAS removal, but most of these materials are faced with high cost or/and low efficiency. The use of metal-organic frameworks (MOFs) as sorbents is attractive for efficient removal of PFAS due to their tailor-made structures and high surface areas. Herein, we synthesized, characterized a water stable Zr-based porphyrinic MOF (PCN-224) with cooperative adsorption domains, and demonstrated its excellent capture performance toward perfluorooctane sulfonate (PFOS), perfluorohexane sulfonate (PFHxS) and perfluorobutane sulfonate (PFBS). PCN-224 has maximum uptake capacities of 963, 517, and 395 mg g-1 for PFOS, PFHxS, and PFBS, respectively, which are much higher than that of granular activated carbon. Moreover, coexistent anions (Cl-, SO4 2-) and humic acid have negligible effects on PFOS adsorption. The excellent adsorption performance of PCN-224 toward PFOS is due to the orthogonal cationic channel pores with a diameter of 1.9 nm, the hydrophobic porphyrin units, and the Zr6 clusters with acidic sites. PCN-224 can be readily regenerated and reused. This work highlights the potential of MOFs with multiple adsorption domains for water remediation.

It takes two: Dual Branch Augmentation Module for domain generalization

Although great success has been achieved in various computer vision tasks, deep neural networks (DNNs) suffer dramatic performance degradation when evaluated on out-of-distribution data. Domain generalization (DG) is proposed to handle this problem by learning domain-agnostic information from multiple source domains to generalize well on unseen target domains. Several methods resort to Fourier transform due to its simplicity and efficiency. They argue that amplitude spectra imply domain-specific information, which should be suppressed, while phase counterparts imply domain-agnostic information, which should be preserved. However, these methods only suppress the domain-specific information in source domains and neglect the relationship with target domains, leading to the persistence of the domain gap. Besides, these methods preserve domain-agnostic information by keeping phase components unchanged, causing them to be underutilized. In this paper, we propose Dual Branch Augmentation Module (DBAM) by leveraging Fourier transform and taking advantage of both amplitude and phase spectra. For the amplitude branch, we propose Inner-domain Amplitude Distribution Rectification (IADR) and Cross-domain Amplitude Dirichlet Mixup (CADM) to stabilize the training process and explore more feature space. In addition, we propose Test-time Amplitude Prototype Calibration (TAPC) to construct the connection between source and target domains during evaluation to further mitigate the domain gap. For the phase branch, we propose Random Symmetric Phase Perturbation (RSPP) to enhance the robustness for recognizing domain-agnostic information. With the combined contributions of the two branches, DBAM significantly surpasses other state-of-the-art (SOTA) methods. Extensive experiments on four benchmarks and further analysis demonstrate the effectiveness of DBAM.

A Privacy-Aware and Incremental Defense Method Against GAN-Based Poisoning Attack

Federated learning is usually utilized as a fraud detection framework in the domain of financial risk management, which promotes the model accuracy without training data exchange. One of the challenges in federated learning is the GAN-based poisoning attack. The GAN-based poisoning attack is a type of intractable poisoning attack that causes global model accuracy degradation and privacy leak. Most of the existing defenses for GAN-based poisoning attack have the three problems: 1) dependence on validation datasets; 2) incompetence of dealing with incremental poisoning attack; and 3) privacy leak. To address the above problems, we present a privacy-aware and incremental defense (PID) method to detect malicious participants and protect privacy. In PID, we design a method to accumulate the offset of model parameters from participants in all current epochs to represent the moving tendency for model parameters. Thus, we can distinguish the adversaries from normal participants based on the accumulations in this incremental poisoning attack. We also use multiple trust domains to reduce the rate of misjudging benign participants as adversaries. Moreover, a differentiated differential privacy is utilized before the global model sending to protect the privacy of participants’ training datasets in PID. The experiments conducted on two real-world datasets under financial fraud detection scenario demonstrate that the PID reduces the fallout of adversaries detection (the rate of misjudging benign participants as adversaries) by at least 51.1% and improve the speed of detecting all malicious participants by at least 33.4% compared with two popular defense methods. Besides, the privacy preserving of PID is also effective.

Ensemble of Domain Adaptation-Based Knowledge Transfer for Evolutionary Multitasking

Recently, a number of domain adaptation methods have been proposed for knowledge transfer in evolutionary multitasking (EMT). However, the learnt mappings in these methods often have unique biases in representing the connection between source and target tasks. Few studies have paid attention to the complementarity of different mappings in knowledge transfer. To fill this research gap, this paper proposes an ensemble method to combine multiple domain adaptation methods for knowledge transfer in EMT by considering the efficacy and diversity of these methods. First, a hierarchical clustering method is used to divide the population of each task into multiple clusters. Then, when two parental solutions are selected for knowledge transfer across tasks, the solutions within the same cluster are checked. In particular, if none of these solutions has been transferred before, the efficacy of domain adaptation methods is considered first by using roulette wheel selection based on the corresponding performance improvements in the evolutionary optimization process. Otherwise, the diversity of domain adaptation methods is emphasized by randomly selecting one of the domain adaptation methods for knowledge transfer. The effectiveness of our proposed ensemble method is validated by embedding it into existing state-of-the-art EMT algorithms, and the experimental results show that our algorithm outperforms several recently proposed EMT algorithms on most cases of two multitasking benchmark suites and one practical case.

When Sparse Graph Representation Learning Falls into Domain Shift: Data Augmentation for Cross-Domain Graph Meta-Learning (Student Abstract)

Cross-domain Graph Meta-learning (CGML) has shown its promise, where meta-knowledge is extracted from few-shot graph data in multiple relevant but distinct domains. However, several recent efforts assume target data available, which commonly does not established in practice. In this paper, we devise a novel Cross-domain Data Augmentation for Graph Meta-Learning (CDA-GML), which incorporates the superiorities of CGML and Data Augmentation， has addressed intractable shortcomings of label sparsity, domain shift, and the absence of target data simultaneously. Specifically, our method simulates instance-level and task-level domain shift to alleviate the cross-domain generalization issue in conventional graph meta-learning. Experiments show that our method outperforms the existing state-of-the-art methods.

Assessing the sustainability of the European Green Deal and its interlin kages with the SDGs

The European Green Deal (EGD) is the growth strategy for Europe, covering multiple domains, and aiming to an equitable, climate neutral European Union by 2050. The UN Agenda 2030, encompassing 17 Sustainable Development Goals (SDGs), establishes the foundation for a global sustainability transition. The integration of the SDGs into the EGD is an overlooked issue in the literature, despite Europe’s slow progress to achieve the sustainability targets. We employed a machine-learning text-mining method to evaluate the extent of SDG integration within the 74 EGD policy documents published during 2019–2023. The findings reveal a substantial alignment of EGD policies with SDGs related to clean energy (SDG7), climate action (SDG13), and sustainable consumption and production (SDG12). In contrast, there is a significant underrepresentation in areas related to social issues such as inequalities, poverty, hunger, health, education, gender equality, decent work, and peace, as indicated by lower alignment with SDGs 1, 2, 3, 4, 5, 8, 10, and 16. Temporal trends suggest a marginal increase in the attention given to environmental health (especially water and marine life) and gender equality. Furthermore, we illustrate the alignment of EGD policies with the six essential sustainability transformations proposed by the Sustainable Development Solutions Network (SDSN) in 2019 for the operationalization of the SDGs. The results indicate that besides the prevalence of “Energy Decarbonization and Sustainable Industry”, all areas have received attention, except for the “Health, Wellbeing and Demography”. The findings call for a more integrated approach to address the complete spectrum of sustainability in a balanced manner.

About them, without them? figures of youth in Australian policy 2014–2021

ABSTRACT This paper considers the multiple and contradictory depictions of young people in youth policy and reflects on how youth engagement is constructed in Australian Federal and State policy. In this journal, researchers have been challenged to critically reflect on ‘figures of youth' (Threadgold, 2020) and we take up this challenge and consider how and why ‘figures of youth' are deployed in Australian youth policy. Drawing on an environmental scan of youth policy documents published between 2014-2021, we analyse the ways young people's voices are positioned to legitimate policy claims about what is required to help young people in Australia across multiple domains. Despite the apparent agreement among policy-makers that youth consultation is important in the policy development process, our scan points to a wide variety of interpretations of what ‘youth voice' might look like, how young people participate, and how their contributions are actually utilised. We argue that the persistence and frequent emergence of particular ‘figures of youth’ throughout Australian youth policy reflects the endurance of adult-centric interpretations of youth, and the partial and superficial procedures that claim to engage with young people, but which ultimately privilege the interests of other agendas.

A Survey of Cutting-edge Multimodal Sentiment Analysis

The rapid growth of the internet has reached the fourth generation, i.e. web 4.0, which supports Sentiment Analysis (SA) in many applications such as social media, marketing, risk management, healthcare, businesses, websites, data mining, e-learning, psychology, and many more. Sentiment analysis is a powerful tool for governments, businesses, and researchers to analyse users’ emotions and mental states in order to generate opinions and reviews about products, services, and daily activities. In the past years, several SA techniques based on Machine Learning (ML), Deep Learning (DL), and other soft computing approaches were proposed. However, growing data size, subjectivity, and diversity pose a significant challenge to enhancing the efficiency of existing techniques and incorporating current development trends, such as Multimodal Sentiment Analysis (MSA) and fusion techniques. With the aim of assisting the enthusiastic researcher to navigating the current trend, this article presents a comprehensive study of various literature to handle different aspects of SA, including current trends and techniques across multiple domains. In order to clarify the future prospects of MSA, this article also highlights open issues and research directions that lead to a number of unresolved challenges.

Time domain coupling of the boundary and discrete element methods for 3D problems

AbstractThis paper presents an extension of the authors’ previously developed interface coupling technique for 2D problems to 3D problems. The method combines the strengths of the Discrete Element Method (DEM), known for its adeptness in capturing discontinuities and non-linearities at the microscale, and the Boundary Element Method (BEM), known for its efficiency in modelling wave propagation within infinite domains. The 3D formulation is based on spherical discrete elements and bilinear quadrilateral boundary elements. The innovative coupling methodology overcomes a critical limitation by enabling the representation of discontinuities within infinite domains, a pivotal development for large-scale dynamic problems. The paper systematically addresses challenges, with a focus on interface compatibility, showcasing the method’s accuracy through benchmark validation on a finite rod and infinite spherical cavity. Finally, a model of a column embedded into the ground illustrates the versatility of the approach in handling complex scenarios with multiple domains. This innovative coupling approach represents a significant leap in the integration of DEM and BEM for 3D problems and opens avenues for tackling complex and realistic problems in various scientific and engineering domains.

How diverse a monocentric chromosome can be? Repeatome and centromeric organization of Juncus effusus (Juncaceae).

Juncus is the largest genus of Juncaceae and was considered holocentric for a long time. Recent findings, however, indicated that 11 species from different clades of the genus have monocentric chromosomes. Thus, the Juncus centromere organization and evolution need to be reassessed. We aimed to investigate the major repetitive DNA sequences of two accessions of Juncus effusus and its centromeric structure by employing whole-genome analyses, fluorescent in situ hybridization, CENH3 immunodetection, and chromatin immunoprecipitation sequencing. We showed that the repetitive fraction of the small J. effusus genome (~270 Mbp/1C) is mainly composed of Class I and Class II transposable elements (TEs) and satellite DNAs. Three identified satellite DNA families were mainly (peri)centromeric, with two being associated with the centromeric protein CENH3, but not strictly centromeric. Two types of centromere organization were discerned in J. effusus: type 1 was characterized by a single CENH3 domain enriched with JefSAT1-155 or JefSAT2-180, whereas type 2 showed multiple CENH3 domains interrupted by other satellites, TEs or genes. Furthermore, while type 1 centromeres showed a higher degree of satellite identity along the array, type 2 centromeres had less homogenized arrays along the multiple CENH3 domains per chromosome. Although the analyses confirmed the monocentric organization of J. effusus chromosomes, our data indicate a more dynamic arrangement of J. effusus centromeres than observed for other plant species, suggesting it may constitute a transient state between mono- and holocentricity.

Improved Regression Analysis with Ensemble Pipeline Approach for Applications across Multiple Domains

In this research, we introduce two new machine learning regression methods: the Ensemble Average and the Pipelined Model. These methods aim to enhance traditional regression analysis for predictive tasks and have undergone thorough evaluation across three datasets, Kaggle House Price, Boston House Price, and California Housing, using various performance metrics. The results consistently show that our models outperform existing methods in terms of accuracy and reliability across all three datasets. The Pipelined Model, in particular, is notable for its ability to combine predictions from multiple models, leading to higher accuracy and impressive scalability. This scalability allows for their application in diverse fields like technology, finance, and healthcare. Furthermore, these models can be adapted for real-time and streaming data analysis, making them valuable for applications such as fraud detection, stock market prediction, and IoT sensor data analysis. Enhancements to the models also make them suitable for big data applications, ensuring their relevance for large datasets and distributed computing environments. It is important to acknowledge some limitations of our models, including potential data biases, specific assumptions, increased complexity, and challenges related to interpretability when using them in practical scenarios. Nevertheless, these innovations advance predictive modeling, and our comprehensive evaluation underscores their potential to provide increased accuracy and reliability across a wide range of applications. The results indicate that the proposed models outperform existing models in terms of accuracy and robustness for all three datasets. The source code can be found at https://huggingface.co/DebajyotyBanik/Ensemble-Pipelined-Regression/tree/main

Canadian intensive care unit nurses' responses to moral distress during the COVID-19 pandemic, and their recommendations for mitigative interventions.

To describe intensive care unit nurses' experiences of moral distress during the COVID-19 pandemic, and their recommendations for mitigative interventions. Interpretive description. Data were collected with a purposeful sample of 40 Canadian intensive care unit nurses between May and September 2021. Nurses completed a demographic questionnaire, the Measure of Moral Distress-Healthcare Professionals survey and in-depth interviews. Quantitative data were analysed using descriptive statistics. Qualitative data were categorized and synthesized using reflexive thematic analysis and rapid qualitative analysis. Half of the nurses in this sample reported moderate levels of moral distress. In response to moral distress, nurses experienced immediate and long-term effects across multiple health domains. To cope, nurses discussed varied reactions, including action, avoidance and acquiescence. Nurses provided recommendations for interventions across multiple organizations to mitigate moral distress and negative health outcomes. Nurses reported that moral distress drove negative health outcomes and attrition in response to moral events in practice. To change these conditions of moral distress, nurses require organizational investments in interventions and cultures that prioritize the inclusion of nursing perspectives and voices. Nurses engage in a variety of responses to cope with moral distress. They possess valuable insights into the practice issues central to moral distress that have significant implications for all members of the healthcare teams, patients and systems. It is essential that nurses' voices be included in the development of future interventions central to the responses to moral distress. This study adheres to COREQ guidelines. What Problem did the Study Address? Given the known structural, systemic and environmental factors that contribute to intensive care unit nurses' experiences of moral distress, and ultimately burnout and attrition, it was important to learn about their experiences of moral distress and their recommendations for organizational mitigative interventions. Documentation of these experiences and recommendations took on a greater urgency during the context of a global health emergency, the COVID-19 pandemic, where such contextual influences on moral distress were less understood. What Were the Main Findings? Over half of the nurses reported a moderate level of moral distress. Nurses who were considering leaving nursing practice reported higher moral distress scores than those who were not considering leaving. In response to moral distress, nurses experienced a variety of outcomes across several health domains. To cope with moral distress, nurses engaged in patterns of action, avoidance and acquiescence. To change the conditions of moral distress, nurses desire organizational interventions, practices and culture changes situated in the amplification of their voices. Where and on Whom Will the Research Have an Impact on? These findings will be of interest to: (1) researchers developing and evaluating interventions that address the complex phenomenon of moral distress, (2) leaders and administrators in hospitals, and relevant healthcare and nursing organizations, and (3) nurses interested in leveraging evidence-informed recommendations to advocate for interventions to address moral distress. What Does this Paper Contribute to the Wider Global Community? This paper advances the body of scientific work on nurses' experiences of moral distress, capturing this phenomenon within the unique context of a global health emergency. Nurses' levels of moral distress using Measure of Moral Distress-Healthcare Professional survey were reported, serving as a comparator for future studies seeking to measure and evaluate intensive care unit nurses' levels of moral distress. Nurses' recommendations for mitigative interventions for moral distress have been reported, which can help inform future interventional studies. No patient or public contribution.

Multivalent coiled-coil interactions enable full-scale centrosome assembly and strength.

The outermost layer of centrosomes, called pericentriolar material (PCM), organizes microtubules for mitotic spindle assembly. The molecular interactions that enable PCM to assemble and resist external forces are poorly understood. Here, we use crosslinking mass spectrometry (XL-MS) to analyze PLK-1-potentiated multimerization of SPD-5, the main PCM scaffold protein in C. elegans. In the unassembled state, SPD-5 exhibits numerous intramolecular crosslinks that are eliminated after phosphorylation by PLK-1. Thus, phosphorylation induces a structural opening of SPD-5 that primes it for assembly. Multimerization of SPD-5 is driven by interactions between multiple dispersed coiled-coil domains. Structural analyses of a phosphorylated region (PReM) in SPD-5 revealed a helical hairpin that dimerizes to form a tetrameric coiled-coil. Mutations within this structure and other interacting regions cause PCM assembly defects that are partly rescued by eliminating microtubule-mediated forces, revealing that PCM assembly and strength are interdependent. We propose that PCM size and strength emerge from specific, multivalent coiled-coil interactions between SPD-5 proteins.

Common mental health conditions and considerations in pediatric chronic kidney disease.

The mental health of youth with chronic kidney disease (CKD) has been increasingly recognized as an area of clinical need. The development of mental health concerns is influenced by a range of physiological, psychological, and environmental factors. Some of these factors are common across child development, but some are more unique to youth with CKD. Mental health concerns are associated with increased risk for a range of poor medical outcomes (e.g., adherence, risk of transplant rejection) and quality of life concerns. In this educational review, we discuss the current evidence base regarding the development of mental health concerns in youth with CKD. The review covers multiple domains including mood and anxiety disorders, traumatic stress, and neurodevelopmental disorders. Estimated prevalence and hypothesized risk factors are outlined, and the potential impact of mental health on medical care and functional outcomes are reviewed. Finally, we introduce options for intervention to support positive mental health and offer recommendations for building access to mental health care and improving the mental health education/training of medical professionals.

Artificial Intelligence: Opportunities and Challenges for the Global South with a Focus on Nepal

Artificial intelligence is a relatively new and contemporary technological tool that also figures significantly in the current discourse of international relations. AI stems from its ability to mimic human cognitive functions, allowing machines to perform tasks that previously required human intelligence. AI systems can analyze massive amounts of data, identify patterns, and make data-driven decisions by utilizing advanced algorithms, machine learning, and data processing capabilities. The primary goal is to improve efficiency, accuracy, and automation across multiple domains, including healthcare, finance, manufacturing, and transportation. Furthermore, AI has the potential to address complex challenges, innovate solutions, and contribute to scientific advancements, making it an effective tool for supplementing human capabilities and driving progress in a variety of fields. However, ethical considerations and responsible development are essential for ensuring the responsible and beneficial integration of AI into society. Concerns regarding AI have also been expressed from the viewpoint of foreign policy, national security, data privacy, and the possible replacement of humans by machines. This is expected to pose additional difficulties for developing nations like Nepal as it may be very expensive to use AI in various development endeavors that aim to address global issues on an equal footing with developed nations. Because most of the countries in the Global South are developing or least developed, AI may therefore present more challenges than opportunities to them.
 Hence, this article attempts to incorporate the various dimensions of AI in terms of the different perspectives that are encountered in this changing context of the AI world, as well as the real impacts that may be opportunities or threats to Nepal from the perspective of international relations and diplomacy.

Applying the NIMHD Research Framework to Social Determinants of Health in the Context of Sport-Related Concussion: A Clinical Commentary.

Sport-related concussion (SRC) is a prevalent injury. There are significant disparities in SRC outcomes across racial and ethnic groups. These disparities may be attributed to the unequal distribution of political power (or influence) and resource allocation in various communities shaping individuals' social determinants of health (SDoH). However, the influence of SDoH on SRC outcomes remains understudied. In this clinical commentary we use the National Institute of Minority Health and Health Disparities Research Framework and describe how it's application can help address gaps in our understanding of SDoH and SRC. This framework provides a comprehensive approach to investigating and addressing health disparities by considering SDoH along multiple levels and domains of influence. Using this framework, athletic trainers (ATs) can identify areas requiring intervention and better understand how SDoH influence SRC outcomes. This understanding can help ATs develop tailored interventions to promote equitable care for SRC.

Applying self-attention model to learn both Empirical Risk Minimization and Invariant Risk Minimization for multimedia recommendation

Multimedia recommendation systems have many applications in our daily life. However, how accurately capture a customer's preference is an issue that is difficult to deal with. The proposed Invariant Risk Minimization (IRM) and Empirical Risk Minimization (ERM) are ways to learn a customer's preference. Still, both frameworks show some limitations: although ERM performs excellently in a single environment, it fails to generalize well when faced with multiple and new domains. On the other hand, IRM learns invariant features across heterogeneous environments, but it lacks theoretical guarantees and performs less effectively where the invariants are unclear. This paper proposes an ERM and IRM Optimized Rating Framework (EIOR) as our final recommender model with direct rating scores. The EIOR enhances the accuracy and functionality of the multimedia recommendation systems by utilizing self-attention mechanisms to combine IRM and ERM with adjusted attention weights. Specifically, IRM learns invariant parts across different environments, while ERM learns variant parts. With self-attention, we can adaptively allocate attention weights for the two pieces and seek the optimal pair of attention weights based on the loss function. We demonstrate EIOR on a cutting-edge recommender model UltraGCN and use the open multimedia dataset of TikTok to finish all the experiments. The results validate the effectiveness of EIOR by comparing purely operating invariant representations alone with the framework of IRM.

Determinants of Change in Physical Activity in Children During the Transition from Elementary to High School.

Effective public health interventions targeting factors that influence physical activity are urgently needed to reduce the age-related decline in physical activity in youth. The purpose of this study was to identify associations between physical activity and a set of potential influences on physical activity in children as they transition from elementary to high school. Participants were 951 children from South Carolina school districts who completed outcome and independent variable measures on at least two time points from the 5th through 11th grades in 2010 - 2017. The primary outcome variable was physical activity, measured by accelerometry. Independent variables included a comprehensive set of variables in the child, parent/home, school, and community domains. Children, parents and school administrators, and staff completed questionnaires to assess psychosocial and home, school, and neighborhood environmental influences. Growth curve analyses identified independent variables associated with physical activity over time, either as a main effect or as an interaction with age. As main effects, self-efficacy, self-schema, sport participation, weekday outdoor hours, importance of child participating in sports and physical activity, safe to play outside, and Physical Activity Resource Assessment weighted score were positively associated with physical activity. The associations between physical activity and enjoyment motivation, appearance motivation, weekend outdoor time, and home equipment exhibited significant interactions with age. Enjoyment motivation influenced physical activity during the earlier years, whereas the remaining three variables influenced physical activity in the later years. Interventions should target multiple domains of influences that may vary by age.

Multi-source unsupervised domain adaptive mill load forecasting method based on deep learning and fusion features

In the grinding industry, accurate prediction of the mill load is the key to enhancing plant profitability and reducing mill failures. After the condition changes, the model is usually mismatched, and the true label cannot be obtained in time. To solve this problem, we propose a multi-source unsupervised domain adaptation model based on deep learning and fusion features (MUDA_DF), which is suitable for predicting multiple mill load parameters. Considering that the information contained in the data of a single working condition is insufficient to completely cover the information of the unmodeled working condition, we introduce multiple source domains into the model, which can improve the generalization of the model. Since the special features of the source domain also contain useful information, we use the features after the fusion of special features and common features for regression prediction. By designing multiple loss functions, we achieve the prediction of three mill load parameters in unknown operating conditions: material-to-ball volume ratio (MBVR), pulp density (PD), and charge volume ratio (CVR). Experiments were carried out on data sets collected by a laboratory ball mill. Compared with the traditional mill load soft measurement method, the R2 of this method can reach above 0.9 and the highest can reach 0.99, which proves the effectiveness of this method.