Knowledge-based Model Research Articles

Fuzzy-UNet: Semantic-Fuzzy Integration for Fabric Defect Identification

Abstract The detection of fabric surface defects is a critical component for quality assurance and operational efficiency within the textile industry. A significant challenge in this field is the effective incorporation of domain expertise to enhance the precision of defect recognition. This paper introduces Fuzzy-UNet, an innovative fabric defect detection system that integrates a semantic segmentation network with a fuzzy decision-making model, addressing the challenge of domain knowledge integration. By combining a cascaded data model with a knowledge-based model, Fuzzy-UNet harnesses the power of deep learning and artificial experience to refine the identification process. The data-driven model of our system is an advanced UNet architecture, specifically tailored for detecting subtle defects in fabrics with non-standard aspect ratios. The system’s novel fuzzy decision model utilizes spatiotemporal data from multiple cameras, which is essential for enhancing the accuracy of the detection process. Our comprehensive experiments demonstrate the robustness of Fuzzy-UNet, with a significant increase in accuracy to 96.10% and a marked reduction in the False Positive Rate (FPR). Fuzzy-UNet's superior performance over existing methods makes it a leading solution for industrial fabric defect detection.

The intrinsic problems in EQ damage prognosis: the experience from a systematic evaluation of L’Aquila databases and different levels of data availability

AbstractWithin the European TURNkey project, a knowledge-based exposure-modelling framework was developed, enabling the consideration of different levels of investigation and data availability. In particular, the proposed framework recognizes various levels and origins of uncertainties, as well as the completeness of a building stock catalogue. Despite substantial efforts, the main question still needs to be answered: How reliable can the developed tools and instruments be if they are not tested and validated by actual events? The L’Aquila 2009 earthquake has been the subject of several analytical strategies to enrich earthquake engineering knowledge. In this study, the information provided by the Italian Observed Damage Database is analyzed, explicitly focusing on the seismic sequence of the L’Aquila 2009 earthquake within the delimited area of the city’s historical center. A second dataset, where the European Macroseismic Scale (EMS-98) was used as a reference, is integrated into the study, and the results are compared. A methodology is implemented for a systematically evaluating the database based on the EMS-98. From the data analysis, a proposal is made to define a comparable EMS-98 building typology and to assign vulnerability classes considering optimistic, pessimistic and most likely criteria. The reliability of the sample is then explored using the knowledge-based exposure modelling framework established by the TURNkey Project. Accuracy is then evaluated through an empirical inspection of frontal (lateral) views available in Google Street View (2022). Images before and after the event are collected and compared with the available data. Intrinsic problems encountered during the process are then listed and discussed, particularly regarding the use of the database, the joint between the studied datasets, and the post-processing required to use the data for damage prognosis. This paper intends to demonstrate how reliable datasets for the building stock, including structural types and corresponding vulnerability classes, can be elaborated. Not least, exposure modelling has to transform the available data into a descriptive form that can be linked directly with the Fragility or Vulnerability Functions, expecting that these assignments are the best suited or representative ones. The data layers provided by the study enable the testing and training of exposure modelling techniques for the selected event and target region.

Enhancing Estimation Accuracy of Prostate Cancer VMAT Planning: A Knowledge-based Approach Using Multiple Collimator Angles.

This study aimed to determine whether a knowledge-based planning model incorporating treatment plans with multiple collimator angles (Multi-coll. model) provides superior estimation accuracy and plan quality for a range of collimator angles compared with a typical institutional model (Inst. model). Five institutions using volumetric modulated arc therapy for prostate cancer participated in the study. The Inst. model comprised plans using a fixed collimator angle, whereas the Multi-coll. model registered plans using single arcs and collimator angles of 10°, 30°, 50°, and 70°. The coefficient of determination and mean squared error were calculated, and the estimation accuracy and dosimetric results for three validation cases at each institution were compared for each model. Dosimetric parameters included volumes receiving at least 20%, 50%, and 90% of the prescription dose (V20, V50, and V90, respectively) at each collimator angle. The Multi-coll. model yielded a higher coefficient of determination and lower mean squared error than the Inst. model, although overfitting of data was a concern at two institutions. The mean absolute errors between estimated and calculated values were (Inst. model vs. Multi-coll. model): 4.41% vs. 2.60% (V20), 4.03% vs. 2.27% (V50), and 2.07% vs. 1.09% (V90) for the rectum; 3.48% vs. 3.42% (V20), 2.24% vs. 3.65% (V50), and 1.11% vs. 0.73% (V90) for the bladder. The Multi-coll. model exhibited a lower rectal and vesical V20 than the Inst. The Multi-coll. model had a greater estimation accuracy and slightly higher plan quality than the Inst.

Optimal planning for hybrid renewable energy systems under limited information based on uncertainty quantification

Developing hybrid renewable energy systems (HRES) is crucial for ensuring affordable, sustainable, and reliable energy access in remote rural and island regions. Rich and detailed data play a pivotal role in the optimal design of HRES. However, remote areas often face data limitations due to inadequate metering infrastructure, which further increases uncertainties in HRES design. This paper proposes a planning optimization method that quantifies multiple uncertainties, requiring only limited information for the optimal HRES design. First, uncertainties in HRES planning are classified into different types based on intrinsic randomness, data availability, and impact level. Then, by integrating uncertainty quantification with credibility theory, knowledge-based scenario generation models are developed to leverage limited available information to generate extreme and typical operating conditions for a tailored HRES design. Finally, a fuzzy credibility-constrained stochastic-robust optimization model is proposed to optimize the design of HRES. A practical case facing significant data scarcity and gaps is used to validate the effectiveness of the proposed method. The proposed approach reduces the loss of load probability for HRES to 0.02% and lowers the levelized cost of energy to $0.16/kWh—33% lower than that of the deterministic method and 79% lower than that of the traditional stochastic method.

Automated optimisation design of barge structures using integrated knowledge-based approaches

ABSTRACT An integrated automatic optimisation simulation system for barge structure design has been developed to enhance efficiency and quality. This system incorporates a knowledge-based geometrical modelling method for barge structure design, allowing for the automatic construction of computational models of barge sections through direct transfer of geometric models. Furthermore, a barge structure optimisation simulation system integrated with direct calculation has been established. The effectiveness of this approach is demonstrated through the successful application of structural optimisation to a 65 m barge section, resulting in higher quality design solutions. This research offers an efficient and effective method for achieving high-quality design solutions in the field of barge structure design.

Structure-mutagenicity relationships on quinoline and indole analogues in the Ames test

BackgroundAlthough the in silico predictive ability of the Ames test results has recently made remarkable progress, there are still some chemical classes for which the predictive ability is not yet sufficient due to a lack of Ames test data. These classes include simple heterocyclic compounds. This study aimed to investigate the mutagenicity and structure-mutagenicity relationships for some heterocycles in the Ames test. In the present study, we selected 12 quinoline analogues containing one or two nitrogen atoms in the naphthalene ring and 12 indole analogues containing one to three nitrogen atoms in the indole ring, without any side moiety.ResultsThe Ames test was performed with five standard bacterial strains (TA100, TA1535, TA98, TA1537, and WP2uvrA) using the pre-incubation method with and without rat liver S9. Five quinoline and two indole analogues were mutagenic. Among the five quinoline analogues, four were mutagenic in the presence of S9 mix with TA100, whereas cinnoline was mutagenic in the absence of S9 mix with TA1537. Among the two indole analogues, indazole was mutagenic in the presence and absence of S9 mix with WP2uvrA and 4-azaindole was mutagenic in the absence of S9 mix with TA1537. The mechanisms underlying the induction of mutagenesis appear to differ between quinoline and indole analogues. In addition, we performed in silico analysis of the mutagenicity of all these analogues using DEREK Nexus 6.1.1 (Lhasa Limited) and GT_EXPERT from CASE Ultra 1.8.0.5 (MultiCASE Inc.) as knowledge-based models and GT1_BMUT from CASE Ultra 1.8.0.5 (MultiCASE Inc.) as a statistical-based model. The knowledge-based model showed low sensitivity for both the quinoline and indole analogues (DEREK Nexus and GT_EXPERT: 20% for quinolines and 0% for indoles). Conversely, the statistical model showed high sensitivity (100% for both quinolines and indoles) and low specificity (43% for quinolines and 10% for indoles).ConclusionBased on the Ames test results, we proposed structural alerts noting that quinoline analogues were mutagenic when they had nitrogens in any of the positions 2, 5, 7, or 8 in addition to 1, and indole analogues were mutagenic when they had nitrogens at positions 2 or 4 in addition to 1.

Data-driven homogenisation of viscoelastic porous elastomers: Feedforward versus knowledge-based neural networks

A computational framework is established to implement time-dependent data-driven surrogate constitutive models for the homogenised mechanical response of porous elastomers at large strains. The aim is to enhance the computational efficiency of multiscale analyses through the use of these surrogate models. To achieve this, explicit finite element (FE) simulations are conducted to predict the homogenised response of a cubic unit cell of a porous elastomer, using two different viscoelastic descriptions of the parent material, subject to pseudo-random, multiaxial, non-proportional histories of macroscopic strains. The histories of homogenised variables extracted from each set of FE predictions form a training dataset, which is used to train two different surrogate models, both relying on artificial neural networks (NNs). The first model predicts the increment in macroscopic stress over a simulation step, as a function of the macroscopic stress and strain at the beginning of the step, of the prescribed macroscopic strain increment, and of the corresponding time increment. The second model uses the same inputs and outputs but tests a knowledge-based modelling approach: it relies on the aid of an additional nonlinear elastic constitutive model, which is time- and path-independent and known a priori. This model represents an existing base of knowledge which is augmented and corrected by a NN after training on viscoelastic data. The data-driven surrogate model, therefore, learns the viscoelastic behaviour of the unit cell starting from knowledge of its elastic response. The two surrogate models are found to have comparable and very high accuracies, capturing the response of the homogenised unit cell to complex loading histories. Hyperparameter optimisation shows that the second, knowledge-based model requires a simpler NN and therefore incurs a smaller computational cost.

Abstract 4139141: Predicted efficacy of inclisiran on cardiovascular outcomes in lower extremity peripheral artery disease: results of the In Silico Sirius study

Introduction: Symptomatic lower extremity peripheral artery disease (PAD) patients have the highest risk of cardiovascular (CV) outcomes among patients with established atherosclerotic cardiovascular disease (ASCVD). Inclisiran, an siRNA targeting PCSK9 mRNA, reduces LDL-C levels. In SIRIUS in silico trial (NCT05974345), inclisiran was predicted to lower CV events in ASCVD patients. Research question/Hypothesis: SIRIUS in silico study aims to predict the efficacy of inclisiran on CV outcomes in subgroups of patients with or without PAD. Methods: The SIRIUS in silico trial was conducted using a knowledge-based mechanistic computational model of ASCVD applied to a virtual ASCVD population with LDL-C ≥ 70 mg/dL. Each virtual patient is its own control. This model was previously calibrated and validated before running the trial. SIRIUS compared the efficacy of inclisiran vs placebo on top of High Intensity (HI) statins with or without ezetimibe on 3-Point-MACE defined as a composite of time to first occurrence of CV death, nonfatal myocardial infarction (MI) or nonfatal ischemic stroke (IS) over 5 years. Occurrence of major acute limb events (MALE) was also individually assessed in time-to-first-event analyses. Results: Among 204, 691 virtual SIRIUS ASCVD patients, 28,072 (13.7%) had PAD. At 5 years, the mean predicted percentage reduction in LDL-C with inclisiran as compared to placebo was -49.3% and -49.8% in patients with or without PAD, respectively. The predicted rate of 3P-MACE in the inclisiran arm was consistently lower than in the placebo arm in patients with PAD (17.62% vs 22.88%; Hazard Ratio (HR): 0.75 medium uncertainty) and in patients without PAD (10.33% vs 13.64%; HR: 0.75 low uncertainty). Compared to placebo, inclisiran was also predicted to consistently reduce MALE in patients with or without PAD (2.71% vs 4.11%; HR: 0.65 medium uncertainty and 0.19% vs 0.29%; HR: 0.66 high uncertainty respectively). Conclusion: Pending the results of ORION-4 and VICTORION-2-P, this first In Silico trial in virtual PAD patients predicted a potential effect of inclisiran on 3P-MACE and MALE reductions over 5 years follow-up.

Abstract 4134364: Predicted Low-Density Lipoprotein Cholesterol and Cardiovascular Outcomes Lowering With Inclisiran in Patients With or Without stroke: Insights from SIRIUS in silico trial.

Introduction: Inclisiran, an siRNA, targeting PCSK9 mRNA, reduces LDL-c levels. In SIRIUS in silico trial (NCT05974345), inclisiran was predicted to lower cardiovascular (CV) events in virtual patients with atherosclerotic cardiovascular disease (ASCVD). Research question: This analysis predicted the potential efficacy of inclisiran on CV outcomes in virtual patients with or without prior ischemic stroke (IS). Methods: The SIRIUS trial was conducted using a calibrated and validated knowledge-based mechanistic computational model of ASCVD applied to a virtual population with LDL-C ≥ 70 mg/dL. Each virtual patient was its own control. SIRIUS compared the efficacy of inclisiran vs placebo on top of High Intensity (HI) statin with or without ezetimibe on 3-Point-MACE defined as a composite of time to first occurrence of CV death, nonfatal myocardial infarction (MI) or nonfatal IS over 5 years in patients with or without prior IS. Occurrence of fatal and non-fatal (IS) was also individually assessed in time-to-first-event analyses. Results: Among 204,691 virtual SIRIUS ASCVD patients, 39 371 (19%) had prior IS. At 5 years, the predicted mean percentage reduction in LDL-C with inclisiran as compared to placebo was –49.17% and –49.88% in patients with or without prior IS respectively. Patients with prior IS were at higher risk of 3P-MACE than patients without IS both with placebo and inclisiran (17.01% vs 14.41% with placebo and 13.44% vs 10.83% with inclisiran). However, the predicted rate of 3P-MACE in the inclisiran arm was consistently lower than in the placebo arm for both prior IS and no prior IS (HR 0.78 medium uncertainty and 0.74 low uncertainty respectively). Compared to placebo, inclisiran was also predicted to consistently reduce fatal and non-fatal IS in patients with or without prior IS (5.45% vs 7.22%; HR: 0.75 medium uncertainty and 1.87% vs 2.54%; HR: 0.73 medium uncertainty respectively). Conclusion: SIRIUS provides insights into the potential efficacy of inclisiran on CV events suggesting a substantial 3P-MACE and fatal and non-fatal IS reduction in ASCVD patients including those with prior IS, several years before the availability of results from ongoing outcomes trials (ORION-4, VICTORION-2P).

Knowledge Editing for Large Language Models: A Survey

Large Language Models (LLMs) have recently transformed both the academic and industrial landscapes due to their remarkable capacity to understand, analyze, and generate texts based on their vast knowledge and reasoning ability. Nevertheless, one major drawback of LLMs is their substantial computational cost for pre-training due to their unprecedented amounts of parameters. The disadvantage is exacerbated when new knowledge frequently needs to be introduced into the pre-trained model. Therefore, it is imperative to develop effective and efficient techniques to update pre-trained LLMs. Traditional methods encode new knowledge in pre-trained LLMs through direct fine-tuning. However, naively re-training LLMs can be computationally intensive and risks degenerating valuable pre-trained knowledge irrelevant to the update in the model. Recently, Knowledge-based Model Editing (KME), also known as Knowledge Editing or Model Editing , has attracted increasing attention, which aims at precisely modifying the LLMs to incorporate specific knowledge, without negatively influencing other irrelevant knowledge. In this survey, we aim at providing a comprehensive and in-depth overview of recent advances in the field of KME. We first introduce a general formulation of KME to encompass different KME strategies. Afterward, we provide an innovative taxonomy of KME techniques based on how the new knowledge is introduced into pre-trained LLMs, and investigate existing KME strategies while analyzing key insights, advantages, and limitations of methods from each category. Moreover, representative metrics, datasets, and applications of KME are introduced accordingly. Finally, we provide an in-depth analysis regarding the practicality and remaining challenges of KME and suggest promising research directions for further advancement in this field.

Multi-strain modeling of influenza vaccine effectiveness in older adults and its dependence on antigenic distance

Influenza vaccine effectiveness (VE) varies seasonally due to host, virus and vaccine characteristics. To investigate how antigenic matching and dosage impact VE, we developed a mechanistic knowledge-based mathematical model. Immunization with a split vaccine is modeled for exposure to A/H1N1 or A/H3N2 virus strains. The model accounts for cross-reactivity of immune cells elicited during previous immunizations with new antigens. We simulated vaccine effectiveness (sVE) of high dose (HD) versus standard dose (SD) vaccines in the older population, from 2011 to 2022. We find that sVE is highly dependent on antigenic matching and that higher dosage improves immunogenicity, activation and memory formation of immune cells. In alignment with clinical observations, the HD vaccine performs better than the SD vaccine in all simulations, supporting the use of the HD vaccine in the older population. This model could be adapted to predict the impact of alternative virus strain selection on clinical outcomes in future influenza seasons.

Analysis of students’ conceptual change in learning Newton’s third law with an integrated framework of model analysis and knowledge integration

The purpose of this study is to examine how students may hold coexisting and competing conceptual ideas and how students’ uses of such ideas may evolve during the process of conceptual change by combining two complementary methods including model analysis and the conceptual framework model of knowledge integration. Conceptual change is fundamental to science learning, highlighting the processes of students’ transitions from predominantly relying on alternative views to using scientific ones. Recent studies suggest a growing consensus that alternative views and scientific conceptions coexist and compete in students’ learning. However, explicit research on how the states of coexistence and competition evolve during the process of conceptual change is underdeveloped. Building on existing research, the model analysis provides a quantitative framework to represent the different consistent and coexisting/competing states through probabilities of consistent and/or mixed (coexisting) use of incorrect and scientific ideas. Meanwhile, the conceptual framework model of knowledge integration provides content-specific, cognitive explanations for the developmental levels and processes of conceptual change from novice to expert. By combining these two methods, researchers can gain insights into how the coexistence and competition among different conceptual ideas may evolve as conceptual change progresses. To target the research goal, a model-based knowledge integration assessment of Newton’s third law is developed to measure students’ levels of knowledge integration and model states in a single test. Quantitative data were collected from Chinese high school students, supplemented by interviews with a subgroup from the same population. The model analysis results show the heterogeneity of conceptual development and the evolving competition across novice, intermediate, and expert students, as well as across different concept dimensions. The interviews provide cognitive explanations for the observed model states, framed from the lens of knowledge integration. These insights can inform the development of effective assessments and instructional strategies to facilitate favorable conceptual change in student learning. Published by the American Physical Society 2024

A physics knowledge-based surrogate model framework for time-dependent slope deformation: Considering water effect and sliding states

The surrogate model serves as an efficient simulation tool during the slope parameter inversion process. However, the creep constitutive model integrated with dynamic damage evolution poses challenges in development of the required surrogate model. In this study, a novel physics knowledge-based surrogate model framework is proposed. In this framework, a Transformer module is employed to capture strain-driven softening-hardening physical mechanisms. Positional encoding and self-attention are utilized to transform the constitutive parameters associated with shear strain, which are not directly time-related, into intermediate latent features for physical loss calculation. Next, a multi-layer stacked GRU (gated recurrent unit) network is built to provide input interfaces for time-dependent intermediate latent features, hydraulic boundary conditions, and water-rock interaction degradation equations, with static parameters introduced via external fully-connected layers. Finally, a combined loss function is constructed to facilitate the collaborative training of physical and data loss, introducing time-dependent weight adjustments to focus the surrogate model on accurate deformation predictions during critical phases. Based on the deformation of a reservoir bank landslide triggered by impoundment and subsequent restabilization, an elasto-viscoplastic constitutive model that considers water effect and sliding state dependencies is developed to validate the proposed surrogate model framework. The results indicate that the framework exhibits good performance in capturing physical mechanisms and predicting creep behavior, reducing errors by about 30 times compared to baseline models such as GRU and LSTM (long short-term memory), meeting the precision requirements for parameter inversion. Ablation experiments also confirmed the effectiveness of the framework. This framework can also serve as a reference for constructing other creep surrogate models that involve non-time-related across dimensions.

Mechanical unfolding of RNA molecules using a knowledge-based model.

We revisit a coarse-grained model to study the dynamics of ribonucleic acid (RNA). In our model, each nucleotide is replaced by an interaction center located at the center of mass. The interaction between nucleotides is carried out by a series of effective pair potentials obtained from the statistical analysis of 501 RNA molecules of high molecular weight from the Protein Data Bank. In addition to the Watson-Crick interactions, we also include non-canonical interactions, which provide stability to the three-dimensional (3D) structure of the molecule. The resulting knowledge-based interactions for the nucleotides (KIN) model allow us to perform efficient Brownian dynamics simulations under different conditions. First, we simulate the stretch of a set of hairpins at a loading rate similar to the values employed in unfolding experiments near equilibrium using optical tweezers. Additionally, we explore unfolding a set of pseudoknots under conditions farther from equilibrium, namely, at loading rates higher than the experimental equilibrium values. The results of our simulations are compared with those obtained from experimental measurements and theoretical models intended to estimate transition states and activation energies. Our KIN model is able to reproduce the intermediate states observed during mechanical unfolding experiments. Moreover, the results of the KIN model are in good agreement with the measured data.

Assessing Scientific Text Similarity: A Novel Approach Utilizing Non-Negative Matrix Factorization and Bidirectional Encoder Representations from Transformer

The patent serves as a vital component of scientific text, and over time, escalating competition has generated a substantial demand for patent analysis encompassing areas such as company strategy and legal services, necessitating fast, accurate, and easily applicable similarity estimators. At present, conducting natural language processing(NLP) on patent content, including titles, abstracts, etc., can serve as an effective method for estimating similarity. However, the traditional NLP approach has some disadvantages, such as the requirement for a huge amount of labeled data and poor explanation of deep-learning-based model internals, exacerbated by the high compression of patent content. On the other hand, most knowledge-based deep learning models require a vast amount of additional analysis results as training variables in similarity estimation, which are limited due to human participation in the analysis part. Thus, in this research, addressing these challenges, we introduce a novel estimator to enhance the transparency of similarity estimation. This approach integrates a patent’s content with international patent classification (IPC), leveraging bidirectional encoder representations from transformers (BERT), and non-negative matrix factorization (NMF). By integrating these techniques, we aim to improve knowledge discovery transparency in NLP across various IPC dimensions and incorporate more background knowledge into context similarity estimation. The experimental results demonstrate that our model is reliable, explainable, highly accurate, and practically usable.

Enhancing Students’ Knowledge-Based Economy Skills at Sultan Qaboos University

This study examines the readiness of Sultan Qaboos University (SQU) students for the knowledge-based economy, focusing on their awareness, perceptions, and skills. As Oman transitions towards a knowledge-based economic model, understanding and enhancing students’ preparedness has become crucial. This research employs a quantitative approach, utilizing a questionnaire distributed to SQU students across various colleges and academic levels, achieving a 66.04% response rate. The key findings reveal that while many students consider knowledge economy skills very important, some express neutral awareness about these skills. Students ranked creativity and innovation (85.82%) and critical thinking and problem solving (85.11%) as the most essential skills. The study identifies significant relationships between enhancing students’ knowledge-based economy skills (ESKBE) and both awareness of the knowledge-based economy (AWS) and the impact of technology on education and employment (ITEE). Students generally view the knowledge economy positively, with 55% agreeing it will have a positive societal impact. However, they recognize challenges, including resistance to change (53.5% agreement) and limited access to education and training (49.3% agreement). The need for government involvement in addressing these challenges was strongly emphasized (54.3% agreement). Based on these findings, the study recommends that SQU enhance its curriculum to explicitly develop key skills, improve awareness of knowledge economy demands, further integrate technology into educational practices, strengthen industry partnerships, promote lifelong learning initiatives, and engage more actively with policymakers. This paper also provides a unique look at higher education’s role in the changing economy, as seen through graduates. It offers empirical evidence for the necessity of universities to change to serve the knowledge economy and, at the same time, make a case on how complex this transformation is. The results matter not only for SQU but also—in a more general sense, at least as far as knowledge-based economies are concerned—globally with respect to higher education institutions.

Towards cleaner energy: An innovative model to minimize NOx emissions in chemical looping and CO2 capture technologies

The urgent global challenges of climate change and environmental issues have catalyzed the development of CO2-capture-ready technologies incorporating fluidization, such as fluidized bed (FB) and circulating fluidized bed (CFB) combustion under oxy-fuel conditions, chemical looping combustion (CLC), chemical looping with oxygen uncoupling (CLOU), in-situ gasification chemical looping combustion (iG-CLC), and calcium (or carbonate) looping (CaL) systems. While these technologies have the potential to reduce CO2 emissions significantly, the persistent presence of nitrogen oxides (NOx = NO + NO2) as pollutants remains a critical environmental concern.This paper introduces a comprehensive fuzzy logic-based model for predicting NOx emissions (FuzzyNOx model) from solid fuel combustion in fluidized beds of chemical looping systems. This innovative model takes into account a wide range of operating parameters, including fuel type, fuel particle size, fuel moisture content, air/fuel ratio, and temperature. It also explores advanced coal and biomass combustion modes, including air-firing, oxyfuel, iG-CLC, and CLOU. Furthermore, it delves into the intricacies of two distinct facilities: the 5 kWth dual-fluidized bed Chemical-Looping-Combustion of Solid-Fuels (DFB-CLC-SF) facility at Czestochowa University of Technology, Poland, and the calcium looping dual-fluidized bed (CaL DFB) facility at Niigata University, Japan.Bituminous coal, semi-anthracite, and wood chips are utilized as fuels. Moreover, three various oxygen carriers (OCs) for chemical looping combustion were employed, namely ilmenite, copper oxide (60 % wt.) supported by carbonate waste from ore flotation, and copper oxide (60 % wt.) supported by ilmenite (20 % wt.) and fly ash. Bituminous coal, semi-anthracite, and wood chips are utilized as fuels.The developed knowledge-based fuzzyNOx model allows the optimization of operating parameters to reduce NOx emissions from CO2 capture technologies.

Exploiting explicit item–item correlations from knowledge graphs for enhanced sequential recommendation

In recent years, the research of employing knowledge graphs (KGs) in sequential recommendation (SR) has received a lot of attention, since the side information extracted from KGs, especially the information of the correlations between items, indeed helps the SR models achieve better performance. However, many previous KG-based SR models tend to introduce some noise information when learning item embeddings, or insufficiently fuse item–item correlations into their sequential modeling, thus limiting their performance improvements. In this paper, we propose a Distance-Aware Knowledge-based Sequential Recommendation model (DAKSR), which exploits the explicit item–item correlations from KGs to achieve enhanced SR. Specifically, as one critical component in our DAKSR, the distance score matrix (DSM) is first obtained to indicate the correlations between items, and then leveraged in the following three major modules of DAKSR. First, in the Item-Set Embedding layer (ISE) all item embeddings are learned based on DSM, in which the noise information is eliminated effectively. Meanwhile, the Knowledge-Infused Transformer (KIT) incorporates DSM into its attention mechanism to improve the feature extraction. Furthermore, the Knowledge Contrastive Learning module (KCL) also leverages the item–item correlations presented in DSM to generate two credible sequence views, which are used to refine sample representations through a contrastive learning strategy, and thus improve the model’s robustness. Our extensive experiments on three SR benchmarks obviously demonstrate our DAKSR’s superior performance over the state-of-the-art (SOTA) KG-based recommendation models. The implementation of our DAKSR is available at https://github.com/Easonsi/DAKSR for reproducing our experiment results conveniently.

Prediction of multiaxial fatigue life with a data-driven knowledge transfer model

A data-driven model is presented for accurate prediction of multiaxial fatigue life based upon the principle of transfer learning (TL). The Tradaboost framework is explored to adjust the weights of training data from different sources, actuating information transfer from domain knowledge to the data-driven modeling of multiaxial fatigue life. Subsequently, extensive experimental results tested under the proportional and non-proportional circle loadings are collected for model evaluation. The results demonstrate that the proposed model is more accurate than domain knowledge-based, conventional data-driven, and comparable TL-based models, with a low data requirement, showcasing good applicability for multiaxial fatigue life assessment.

A dopaminergic basis of behavioral control.

Both goal-directed and automatic processes shape human behavior, but these processes often conflict. Behavioral control is the decision about which process guides behavior. Despite the importance of behavioral control for adaptive decision-making, its neural mechanisms remain unclear. Critically, it is unknown if there are mechanisms for behavioral control that are distinct from those supporting the formation of goal-relevant knowledge. We performed deep phenotyping of individual dopamine system function by combining multiple PET scans, fMRI, and dopaminergic drug administration in a within-subject, double-blind, placebo-controlled design. Subjects performed a rule-based response time task, with goal-directed and automatic decision-making operationalized as model-based and model-free influences on behavior. We found a double dissociation between two aspects of ventral striatal dopamine physiology: D2/3 receptor availability and dopamine synthesis capacity. Convergent and causal evidence indicated that D2/3 receptors regulate behavioral control by enhancing model-based and blunting model-free influences on behavior but do not affect model-based knowledge formation. In contrast, dopamine synthesis capacity was linked to the formation of model-based knowledge but not behavioral control. D2/3 receptors also modulated frontostriatal functional connectivity, suggesting they regulate behavioral control by gating prefrontal inputs to the striatum. These results identify central mechanisms underlying individual and state differences in behavioral control and point to striatal D2/3 receptors as targets for interventions for improving goal-directed behavior.