Search

In many settings, fleets of assets must perform series of missions with in-between finite breaks. For such fleets, a widely used maintenance strategy is the fleet selective maintenance (FSM). Under resource constraints, the FSM problem selects an optimal subset of feasible maintenance actions to be performed on a subset of components to minimise the maintenance cost while ensuring high system reliability during the upcoming mission. The majority of extant FSMP models are focussed on traditional physics-based reliability models. With recent advances in Machine Learning (ML) and Deep Learning (DL) algorithms, data-driven methods have shown accuracy in predicting remaining useful life (RUL). This paper proposes a predictive FSM strategy for fleets of complex and large multicomponent systems. It relies on a concurrent ML/DL and optimisation approach where a clustering algorithm is used to determine the health states of components and a probabilistic RUL prognostics model is used for component reliability assessment. An optimisation model is developed to solve the predictive FSM problem to ensure high reliability of all systems within the fleet. An efficient two-phase solution approach is developed to solve this complex optimisation problem. Numerical experiments show the validity of the approach and highlight the improved maintenance plans achieved. [ABSTRACT FROM AUTHOR]

The increasing societal concern for consumer information privacy has led to the enforcement of privacy regulations worldwide. In an effort to adhere to privacy regulations such as the General Data Protection Regulation (GDPR), many companies’ privacy policies have become increasingly lengthy and complex. In this study, we adopted the computational design science paradigm to design a novel privacy policy evolution analytics framework to help identify how companies change and present their privacy policies based on privacy regulations. The framework includes a self-attentive annotation system (SAAS) that automatically annotates paragraph-length segments in privacy policies to help stakeholders identify data practices of interest for further investigation. We rigorously evaluated SAAS against state-of-the-art machine learning (ML) and deep learning (DL)-based methods on a well-established privacy policy dataset, OPP-115. SAAS outperformed conventional ML and DL models in terms of F1-score by statistically significant margins. We demonstrate the proposed framework’s practical utility with an in-depth case study of GDPR’s impact on Amazon’s privacy policies. The case study results indicate that Amazon’s post-GDPR privacy policy potentially violates a fundamental principle of GDPR by causing consumers to exert more effort to find information about first-party data collection. Given the increasing importance of consumer information privacy, the proposed framework has important implications for regulators and companies. We discuss several design principles followed by the SAAS that can help guide future design science-based e-commerce, health, and privacy research. [ABSTRACT FROM AUTHOR]

Machine-learning interatomic potential models based on graph neural network architectures have the potential to make atomistic materials modeling widely accessible due to their computational efficiency, scalability, and broad applicability. The training datasets for many such models are derived from density-functional theory calculations, typically using a semilocal exchange-correlation functional. As a result, long-range interactions such as London dispersion are often missing in these models. We investigate whether this missing component can be addressed by combining a graph deep learning potential with semiempirical dispersion models. We assess this combination by deriving the equations of state for layered pnictogen chalcohalides BiTeBr and BiTeI and performing crystal structure optimizations for a broader set of V–VI–VII compounds with various stoichiometries, many of which possess van der Waals gaps. We characterize the optimized crystal structures by calculating their x-ray diffraction patterns and radial distribution function histograms, which are also used to compute Earth mover's distances to quantify the dissimilarity between the optimized and corresponding experimental structures. We find that dispersion-corrected graph deep learning potentials generally (though not universally) provide a more realistic description of these compounds due to the inclusion of van der Waals attractions. In particular, their use results in systematic improvements in predicting not only the van der Waals gap but also the layer thickness in layered V–VI–VII compounds. Our results demonstrate that the combined potentials studied here, derived from a straightforward approach that neither requires fine-tuning the training nor refitting the potential parameters, can significantly improve the description of layered polar crystals. [ABSTRACT FROM AUTHOR]

We uncover the origin of unique electronic structures of single-atom alloys (SAAs) by interpretable deep learning. The approach integrates tight-binding moment theory with graph neural networks to accurately describe the local electronic structure of transition and noble metal sites upon perturbation. We emphasize the complex interplay of interatomic orbital coupling and on-site orbital resonance, which shapes the d-band characteristics of an active site, shedding light on the origin of free-atom-like d-states that are often observed in SAAs involving d10 metal hosts. This theory-infused neural network approach significantly enhances our understanding of the electronic properties of single-site catalytic materials beyond traditional theories. [ABSTRACT FROM AUTHOR]

A deep learning neural network-assisted design strategy for programmable piezoelectric phononic crystal (PnC) beams with shunt circuits is proposed. The feasibility of integrating deep learning into the design of tunable PnCs to achieve real-time vibration isolation is demonstrated through numerical examples. The influence of shunt circuits (capacitance) on bandgaps of piezoelectric PnCs is studied by finite element (FE) simulations. The results show that the bandgap frequency and range vary with the capacitance and electrode length. Moreover, incorporating supercell structures introduces an additional bandgap, significantly expanding the tunable range of the bandgap and demonstrating that shunt circuit modifications can tailor the frequency and width of the bandgap. A suite of deep learning neural network (NN) algorithms is developed for predicting bandgaps and inversely designing PnC parameters, greatly accelerating the bandgap calculation and enabling faster inverse design than existing models. The accuracy of the NN algorithms is verified by comparing their predictions with those from FE simulations. The combination of designed PnC beams and deep learning NNs enables real-time vibration reduction and isolation. This design strategy is successfully validated in a practical scenario involving real-time vibration isolation of train rails. [ABSTRACT FROM AUTHOR]

Small angle x-ray scattering (SAXS) is a widely recognized solution for measuring complex nanostructures. With the increasing demand for accurately assessing structural characteristics and optimizing manufacturing processes, uncertainty quantification in SAXS inverse problems has become a critical issue. However, traditional methods face challenges such as slow computation speed and inaccurate estimation of multidimensional parameters. To overcome these issues, we propose an uncertainty quantification approach suitable for SAXS measurement that approximates the posterior using Bayesian deep learning. The effectiveness and reliability of our method are illustrated by assessing structural parameters of synthetic 2D Si grating samples. The uncertainty quantification takes only about 2.3 s, thousands of times faster than the conventional Markov Chain Monte Carlo (MCMC) methods. Also, our method has superior repeatability for parameter measurement compared to the MCMC approaches. It provides the potential of efficient and reliable SAXS measurement in increasingly intricate semiconductor manufacturing. [ABSTRACT FROM AUTHOR]

The contribution of nuclear quantum effects (NQEs) to the properties of various hydrogen-bound systems, including biomolecules, is increasingly recognized. Despite the development of many acceleration techniques, the computational overhead of incorporating NQEs in complex systems is sizable, particularly at low temperatures. In this work, we leverage deep learning and multiscale coarse-graining techniques to mitigate the computational burden of path integral molecular dynamics (PIMD). In particular, we employ a machine-learned potential to accurately represent corrections to classical potentials, thereby significantly reducing the computational cost of simulating NQEs. We validate our approach using four distinct systems: Morse potential, Zundel cation, single water molecule, and bulk water. Our framework allows us to accurately compute position-dependent static properties, as demonstrated by the excellent agreement obtained between the machine-learned potential and computationally intensive PIMD calculations, even in the presence of strong NQEs. This approach opens the way to the development of transferable machine-learned potentials capable of accurately reproducing NQEs in a wide range of molecular systems. [ABSTRACT FROM AUTHOR]

The transition from a highly subjective morphological assessment to time-lapse imaging improves the accuracy of predicting embryonic developmental potential. In actual operations, embryos are cultured for 2–3 days in a time-lapse monitoring system before being transferred to recipients. However, most existing prediction models require videos or images spanning a five-day period. Therefore, it is necessary to develop a method that accurately predicts blastocyst formation given input data spanning only 2–3 days. In this study, we propose a method for predicting blastocyst formation using early morphokinetic and morphological parameters prior to the five-cell stage. We employed a YOLOv5 pretrained deep-learning network to recognize the first four-cell stages for the accurate extraction of morphokinetic parameters and used these parameters as inputs to construct four long short-term memory-based morphokinetic models for blastocyst formation prediction, obtaining the best area-under-the-curve (AUC) value of 0.7297 [0.669–0.884]. We then extracted the three frames before and after the t1–t4 time points and calculated the image entropy and gray-level co-occurrence matrix entropy as morphological features to build a prediction model. This model was subsequently fused with the morphokinetic model, and an AUC of 0.8325 [0.7601–0.9067] was achieved. Our results have implications for automatic embryo screening given information on early embryonic development. [ABSTRACT FROM AUTHOR]

Anomaly detection is a frequently discussed topic in manufacturing. However, the issues of anomaly detection are typically attributed to the manufacturing process or equipment itself. In practice, the sensor responsible for collecting data and monitoring values may fail, leading to a biased detection result – false alarm. In such cases, replacing the sensor is necessary instead of performing equipment maintenance. This study proposes an effective framework embedded with autoencoder-based control limits that can dynamically distinguish sensor anomaly from process anomaly in real-time. We conduct a simulation numerical study and an empirical study of semiconductor assembling manufacturers to validate the proposed framework. The results show that the proposed model outperforms other benchmark methods and can successfully identify sensor failures, even under conditions of (1) large variations in process values or sensor values and (2) heteroscedasticity effect. This is particularly beneficial in various practical applications where sensors are used for numerical measurements and support equipment maintenance. [ABSTRACT FROM AUTHOR]

This study conducts a comprehensive analysis of deep reinforcement learning (DRL) algorithms applied to supply chain inventory management (SCIM), which consists of a sequential decision-making problem focussed on determining the optimal quantity of products to produce and ship across multiple capacitated local warehouses over a specific time horizon. In detail, we formulate the problem as a Markov decision process for a divergent two-echelon inventory control system characterised by stochastic and seasonal demand, also presenting a balanced allocation rule designed to prevent backorders in the first echelon. Through numerical experiments, we evaluate the performance of state-of-the-art DRL algorithms and static inventory policies in terms of both cost minimisation and training time while varying the number of local warehouses and product types and the length of replenishment lead times. Our results reveal that the Proximal Policy Optimization algorithm consistently outperforms other algorithms across all experiments, proving to be a robust method for tackling the SCIM problem. Furthermore, the (s, Q)-policy stands as a solid alternative, offering a compromise in performance and computational efficiency. Lastly, this study presents an open-source software library that provides a customisable simulation environment for addressing the SCIM problem, utilising a wide range of DRL algorithms and static inventory policies. [ABSTRACT FROM AUTHOR]

The Journal of Chemical Physics has released a special issue focused on software for atomistic machine learning. This issue aims to address the lack of journals dedicated to publishing scientific software papers. The collection of papers in this issue provides insight into the tools and goals of software implementations in the field of atomistic machine learning. The articles cover a range of topics, including machine-learning interatomic potentials, sampling, dataset repositories, workflows, and auxiliary tooling and analysis. The article concludes by emphasizing the importance of software implementations in the field and encourages further submissions on relevant topics. [Extracted from the article]

Photo modulated optical reflectance (PMOR) is an ideal ultra-shallow junction area metrology technique for measurement of transistor dopant distribution in integrated circuit fabrication, and the characterization of process parameters such as implant energy, implant angle, and implant dose has a significant impact on the accuracy of the ion implantation process. This study utilized deep learning to analyze various process parameters concurrently and assessed its performance on boron-doped silicon samples, the data were obtained from the power curves measured from Carrier Illumination (CI) experiments in PMOR, a deep learning model with multi-task learning architecture was developed and trained to characterize multiple process parameters, and the PMOR model incorporating a multi-task learning architecture for process parameters demonstrated superior performance in terms of accuracy and speed of characterization. The analyses indicated that applying deep learning methods to CI metrology in PMOR technology is feasible. In particular, compared with the conventional carrier irradiation technique, the ability to obtain the implantation dose and doping profile along with other process parameters such as implantation energy, implantation angle, and implantation type can better assist in the accurate realization of the ion implantation process with acceptable accuracy and time cost. [ABSTRACT FROM AUTHOR]

Automatic determination of manufacturing process sequences for the physical production of given part designs is key to facilitate on-demand cyber manufacturing. In this work, we propose an integrated framework that (i) identifies manufacturing features from 3D part designs using a Graph Neural Network (GNN), (ii) identifies the manufacturing processes necessary to produce all features in the part using a Convolutional Neural Network (CNN) that considers shape, material properties, and quality information, and (iii) outputs an ordered manufacturing sequence that can produce the designed part with the help of sequence mining. Using these methods, the knowledge required to enable automated manufacturing process selection is easily scalable and updatable without requiring manual population of ad-hoc or rule-based descriptions. We present exemplar implementations of the proposed framework by suggesting manufacturing sequences for discrete parts with multiple features. The suggested manufacturing sequences demonstrate the potential of the proposed framework for use in future on-demand cyber manufacturing applications. [ABSTRACT FROM AUTHOR]

This paper presents BubbleID, a sophisticated deep learning architecture designed to comprehensively identify both static and dynamic attributes of bubbles within sequences of boiling images. By amalgamating segmentation powered by Mask R-CNN with SORT-based tracking techniques, the framework is capable of analyzing each bubble's location, dimensions, interface shape, and velocity over its lifetime and capturing dynamic events such as bubble departure. BubbleID is trained and tested on boiling images across diverse heater surfaces and operational settings. This paper also offers a comparative analysis of bubble interface dynamics prior to and post-critical heat flux conditions. [ABSTRACT FROM AUTHOR]

This document is an editorial introduction from the Journal of Management Information Systems. It emphasizes the importance of incorporating artificial intelligence (AI) into the field of Information Systems (IS) research. The editorial discusses the potential benefits and risks of AI, including its impact on societal well-being, economic benefits, and long-term threats. It also highlights several research papers published in the journal that explore topics such as trust in AI, AI investments, cybersecurity, value creation through AI systems, and the effects of customization in e-commerce. The editorial encourages the IS community to contribute to the understanding and deployment of AI in order to enhance organizational development and human capabilities. [Extracted from the article]

Cybersecurity is a present and growing concern that needs to be addressed with both behavioral and design-oriented research. Public cloud providers such as Amazon Web Services and federal funding agencies such as the National Science Foundation have invested billions of dollars into developing high-performance computing resources accessible to users through configurable virtual machine (VM) images. This approach offers users the flexibility of changing and updating their environment for their computational needs. Despite the substantial benefits, users often introduce thousands of vulnerabilities by installing open-source software packages and misconfiguring file systems. Given the scale of vulnerabilities, security personnel struggle to identify and prioritize vulnerable assets for remediation. In this research, we designed a novel unsupervised deep learning-based Multi-View Combinatorial-Attentive Autoencoder (MV-CAAE) to capture multi-dimensional vulnerability data and automatically identify groups of similar vulnerable compute instances to help facilitate the development of targeted remediation strategies. We rigorously evaluated the proposed MV-CAAE against state-of-the-art methods in three technical clustering experiments. Experiment results indicate that the MV-CAAE achieves V-measure scores (metric of cluster quality) 8 percent-48 percent higher than benchmark methods. We demonstrated the practical value through a comprehensive case study by clustering vulnerable VMs and gathering qualitative feedback from experienced security professionals through semi-structured interviews. The results indicated that clustering vulnerable assets can help prioritize vulnerable instances for remediation and enhance decision-making tasks. The present design-research work also contributes to our theoretical knowledge of cyber-defense. [ABSTRACT FROM AUTHOR]

This paper considers two pertinent research inquiries: 'Can an AI-based predictive framework be utilised for the optimisation of solar energy management?' and 'What are the ways in which the AI-based predictive framework can be integrated within the Smart Grid infrastructure to improve grid reliability and efficiency?' The study deploys a Deep Learning model based on Long Short-Term Memory techniques, leading to refined accuracy in solar electricity generation forecasts. Such an AI-supported methodology aids power grid operators in comprehensive planning, thereby ensuring a robust electricity supply. The effectiveness of this framework is tested using performance metrics such as MAE, RMSE, nMAE, nRMSE, and $ R^2 $ R 2 . A persistent model is utilised as a reference for comparison. Despite a slight decrease in predictive precision with the expansion of the forecast horizon, the proposed AI-based framework consistently surpasses the persistent model, particularly for horizons beyond two hours. Therefore, this research underscores the potential of AI-based prediction in fostering efficient solar energy management and enhancing Smart Grid reliability and efficiency. [ABSTRACT FROM AUTHOR]

Liquid biopsies have the potential to revolutionize cancer care through non-invasive early detection of tumors. Developing a robust liquid biopsy test requires collecting high-dimensional data from a large number of blood samples across heterogeneous groups of patients. We propose that the generative capability of variational auto-encoders enables learning a robust and generalizable signature of blood-based biomarkers. In this study, we analyze orphan non-coding RNAs (oncRNAs) from serum samples of 1050 individuals diagnosed with non-small cell lung cancer (NSCLC) at various stages, as well as sex-, age-, and BMI-matched controls. We demonstrate that our multi-task generative AI model, Orion, surpasses commonly used methods in both overall performance and generalizability to held-out datasets. Orion achieves an overall sensitivity of 94% (95% CI: 87%-98%) at 87% (95% CI: 81%-93%) specificity for cancer detection across all stages, outperforming the sensitivity of other methods on held-out validation datasets by more than  ~ 30%.

Behavioral larval zebrafish screens leverage a high-throughput small molecule discovery format to find neuroactive molecules relevant to mammalian physiology. We screen a library of 650 central nervous system active compounds in high replicate to train deep metric learning models on zebrafish behavioral profiles. The machine learning initially exploited subtle artifacts in the phenotypic screen, necessitating a complete experimental re-run with rigorous physical well-wise randomization. These large matched phenotypic screening datasets (initial and well-randomized) provide a unique opportunity to quantify and understand shortcut learning in a full-scale, real-world drug discovery dataset. The final deep metric learning model substantially outperforms correlation distance-the canonical way of computing distances between profiles-and generalizes to an orthogonal dataset of diverse drug-like compounds. We validate predictions by prospective in vitro radio-ligand binding assays against human protein targets, achieving a hit rate of 58% despite crossing species and chemical scaffold boundaries. These neuroactive compounds exhibit diverse chemical scaffolds, demonstrating that zebrafish phenotypic screens combined with metric learning achieve robust scaffold hopping capabilities.

CRISPR-Cas systems have transformed the field of synthetic biology by providing a versatile method for genome editing. The efficiency of CRISPR systems is largely dependent on the sequence of the constituent sgRNA, necessitating the development of computational methods for designing active sgRNAs. While deep learning-based models have shown promise in predicting sgRNA activity, the accuracy of prediction is primarily governed by the data set used in model training. Here, we trained a convolutional neural network (CNN) model and a large language model (LLM) on balanced and imbalanced data sets generated from CRISPR-Cas12a screening data for the yeast Yarrowia lipolytica and evaluated their ability to predict high- and low-activity sgRNAs. We further tested whether prediction performance can be improved by training on imbalanced data sets augmented with synthetic sgRNAs. Lastly, we demonstrated that adding synthetic sgRNAs to inherently imbalanced CRISPR-Cas9 data sets from Y. lipolytica and Komagataella phaffii leads to improved performance in predicting sgRNA activity, thus underscoring the importance of employing balanced training sets for accurate sgRNA activity prediction.

PURPOSE: Accurate and rapid needle localization on 3D magnetic resonance imaging (MRI) is critical for MRI-guided percutaneous interventions. The current workflow requires manual needle localization on 3D MRI, which is time-consuming and cumbersome. Automatic methods using 2D deep learning networks for needle segmentation require manual image plane localization, while 3D networks are challenged by the need for sufficient training datasets. This work aimed to develop an automatic deep learning-based pipeline for accurate and rapid 3D needle localization on in vivo intra-procedural 3D MRI using a limited training dataset. METHODS: The proposed automatic pipeline adopted Shifted Window (Swin) Transformers and employed a coarse-to-fine segmentation strategy: (1) initial 3D needle feature segmentation with 3D Swin UNEt TRansfomer (UNETR); (2) generation of a 2D reformatted image containing the needle feature; (3) fine 2D needle feature segmentation with 2D Swin Transformer and calculation of 3D needle tip position and axis orientation. Pre-training and data augmentation were performed to improve network training. The pipeline was evaluated via cross-validation with 49 in vivo intra-procedural 3D MR images from preclinical pig experiments. The needle tip and axis localization errors were compared with human intra-reader variation using the Wilcoxon signed rank test, with p

Precise size and shape control in nanocrystal synthesis is essential for utilizing nanocrystals in various industrial applications, such as catalysis, sensing, and energy conversion. However, traditional ensemble measurements often overlook the subtle size and shape distributions of individual nanocrystals, hindering the establishment of robust structure-property relationships. In this study, we uncover intricate shape evolutions and growth mechanisms in Co3O4 nanocrystal synthesis at a subnanometer scale, enabled by deep-learning-assisted statistical characterization. By first controlling synthetic parameters such as cobalt precursor concentration and water amount then using high resolution electron microscopy imaging to identify the geometric features of individual nanocrystals, this study provides insights into the interplay between synthesis conditions and the size-dependent shape evolution in colloidal nanocrystals. Utilizing population-wide imaging data encompassing over 441,067 nanocrystals, we analyze their characteristics and elucidate previously unobserved size-resolved shape evolution. This high-throughput statistical analysis is essential for representing the entire population accurately and enables the study of the size dependency of growth regimes in shaping nanocrystals. Our findings provide experimental quantification of the growth regime transition based on the size of the crystals, specifically (i) for faceting and (ii) from thermodynamic to kinetic, as evidenced by transitions from convex to concave polyhedral crystals. Additionally, we introduce the concept of an onset radius, which describes the critical size thresholds at which these transitions occur. This discovery has implications beyond achieving nanocrystals with desired morphology; it enables finely tuned correlation between geometry and material properties, advancing the field of colloidal nanocrystal synthesis and its applications.

Successful integration of point-of-care testing (POCT) into clinical settings requires improved assay sensitivity and precision to match laboratory standards. Here, we show how innovations in amplified biosensing, imaging, and data processing, coupled with deep learning, can help improve POCT. To demonstrate the performance of our approach, we present a rapid and cost-effective paper-based high-sensitivity vertical flow assay (hs-VFA) for quantitative measurement of cardiac troponin I (cTnI), a biomarker widely used for measuring acute cardiac damage and assessing cardiovascular risk. The hs-VFA includes a colorimetric paper-based sensor, a portable reader with time-lapse imaging, and computational algorithms for digital assay validation and outlier detection. Operating at the level of a rapid at-home test, the hs-VFA enabled the accurate quantification of cTnI using 50 μL of serum within 15 min per test and achieved a detection limit of 0.2 pg/mL, enabled by gold ion amplification chemistry and time-lapse imaging. It also achieved high precision with a coefficient of variation of

Research on action-based timing has shed light on the temporal dynamics of sensorimotor coordination. This study investigates the neural mechanisms underlying action-based timing, particularly during finger-tapping tasks involving synchronized and syncopated patterns. Twelve healthy participants completed a continuation task, alternating between tapping in time with an auditory metronome (pacing) and continuing without it (continuation). Electroencephalography data were collected to explore how neural activity changes across these coordination modes and phases. We applied deep learning methods to classify single-trial electroencephalography data and predict behavioral timing conditions. Results showed significant classification accuracy for distinguishing between pacing and continuation phases, particularly during the presence of auditory cues, emphasizing the role of auditory input in motor timing. However, when auditory components were removed from the electroencephalography data, the differentiation between phases became inconclusive. Mean accuracy asynchrony, a measure of timing error, emerged as a superior predictor of performance variability compared to inter-response interval. These findings highlight the importance of auditory cues in modulating motor timing behaviors and present the challenges of isolating motor activation in the absence of auditory stimuli. Our study offers new insights into the neural dynamics of motor timing and demonstrates the utility of deep learning in analyzing single-trial electroencephalography data.

BACKGROUND: Aortic valve stenosis (AS) is a progressive chronic disease with progression rates that vary in patients and therefore difficult to predict. OBJECTIVES: The aim of this study was to predict the progression of AS using comprehensive and longitudinal patient data. METHODS: Machine and deep learning algorithms were trained on a data set of 303 patients enrolled in the PROGRESSA (Metabolic Determinants of the Progression of Aortic Stenosis) study who underwent clinical and echocardiographic follow-up on an annual basis. Performance of the models was measured to predict disease progression over long (next 5 years) and short (next 2 years) terms and was compared to a standard clinical model with usually used features in clinical settings based on logistic regression. RESULTS: For each annual follow-up visit including baseline, we trained various supervised learning algorithms in predicting disease progression at 2- and 5-year terms. At both terms, LightGBM consistently outperformed other models with the highest average area under curves across patient visits (0.85 at 2 years, 0.83 at 5 years). Recurrent neural network-based models (Gated Recurrent Unit and Long Short-Term Memory) and XGBoost also demonstrated strong predictive capabilities, while the clinical model showed the lowest performance. CONCLUSIONS: This study demonstrates how an artificial intelligence-guided approach in clinical routine could help enhance risk stratification of AS. It presents models based on multisource comprehensive data to predict disease progression and clinical outcomes in patients with mild-to-moderate AS at baseline.

OBJECTIVES: Serum creatinine (SCr) is the primary biomarker for assessing kidney function; however, it may lag behind true kidney function, especially in instances of acute kidney injury (AKI). The objective of the work is to develop Nephrocast, a deep-learning model to predict next-day SCr in adult patients treated in the intensive care unit (ICU). MATERIALS AND METHODS: Nephrocast was trained and validated, temporally and prospectively, using electronic health record data of adult patients admitted to the ICU in the University of California San Diego Health (UCSDH) between January 1, 2016 and June 22, 2024. The model features consisted of demographics, comorbidities, vital signs and laboratory measurements, and medications. Model performance was evaluated by mean absolute error (MAE) and root-mean-square error (RMSE) and compared against the prediction days SCr as a reference. RESULTS: A total of 28 191 encounters met the eligibility criteria, corresponding to 105 718 patient-days. The median (interquartile range [IQR]) MAE and RMSE in the internal test set were 0.09 (0.085-0.09) mg/dL and 0.15 (0.146-0.152) mg/dL, respectively. In the prospective validation, the MAE and RMSE were 0.09 mg/dL and 0.14 mg/dL, respectively. The models performance was superior to the reference SCr. DISCUSSION AND CONCLUSION: Our model demonstrated good performance in predicting next-day SCr by leveraging clinical data routinely collected in the ICU. The model could aid clinicians in in identifying high-risk patients for AKI, predicting AKI trajectory, and informing the dosing of renally eliminated drugs.

We proposed an end-to-end deep learning convolutional neural network (DCNN) for region-of-interest based multi-parameter quantification (RMQ-Net) to accelerate quantitative ultrashort echo time (UTE) MRI of the knee joint with automatic multi-tissue segmentation and relaxometry mapping. The study involved UTE-based T1 (UTE-T1) and Adiabatic T1ρ (UTE-AdiabT1ρ) mapping of the knee joint of 65 human subjects, including 20 normal controls, 29 with doubtful-minimal osteoarthritis (OA), and 16 with moderate-severe OA. Comparison studies were performed on UTE-T1 and UTE-AdiabT1ρ measurements using 100%, 43%, 26%, and 18% UTE MRI data as the inputs and the effects on the prediction quality of the RMQ-Net. The RMQ-net was modified and retrained accordingly with different combinations of inputs. Both ROI-based and voxel-based Pearson correlation analyses were performed. High Pearson correlation coefficients were achieved between the RMQ-Net predicted UTE-T1 and UTE-AdiabT1ρ results and the ground truth for segmented cartilage with acceleration factors ranging from 2.3 to 5.7. With an acceleration factor of 5.7, the Pearson r-value achieved 0.908 (ROI-based) and 0.945 (voxel-based) for UTE-T1, and 0.733 (ROI-based) and 0.895 (voxel-based) for UTE-AdiabT1ρ, correspondingly. The results demonstrated that RMQ-net can significantly accelerate quantitative UTE imaging with automated segmentation of articular cartilage in the knee joint.

PURPOSE: Current clinical risk stratification methods for localized prostate cancer are suboptimal, leading to over- and undertreatment. Recently, machine learning approaches using digital histopathology have shown superior prognostic ability in phase III trials. This study aims to develop a clinically usable risk grouping system using multimodal artificial intelligence (MMAI) models that outperform current National Comprehensive Cancer Network (NCCN) risk groups. MATERIALS AND METHODS: The cohort comprised 9,787 patients with localized prostate cancer from eight NRG Oncology randomized phase III trials, treated with radiation therapy, androgen deprivation therapy, and/or chemotherapy. Locked MMAI models, which used digital histopathology images and clinical data, were applied to each patient. Expert consensus on cut points defined low-, intermediate-, and high-risk groups on the basis of 10-year distant metastasis rates of 3% and 10%, respectively. The MMAIs reclassification and prognostic performance were compared with the three-tier NCCN risk groups. RESULTS: The median follow-up for censored patients was 7.9 years. According to NCCN risk categories, 30.4% of patients were low-risk, 25.5% intermediate-risk, and 44.1% high-risk. The MMAI risk classification identified 43.5% of patients as low-risk, 34.6% as intermediate-risk, and 21.8% as high-risk. MMAI reclassified 1,039 (42.0%) patients initially categorized by NCCN. Despite the MMAI low-risk group being larger than the NCCN low-risk group, the 10-year metastasis risks were comparable: 1.7% (95% CI, 0.2 to 3.2) for NCCN and 3.2% (95% CI, 1.7 to 4.7) for MMAI. The overall 10-year metastasis risk for NCCN high-risk patients was 16.6%, with MMAI further stratifying this group into low-, intermediate-, and high-risk, showing metastasis rates of 3.4%, 8.2%, and 26.3%, respectively. CONCLUSION: The MMAI risk grouping system expands the population of men identified as having low metastatic risk and accurately pinpoints a high-risk subset with elevated metastasis rates. This approach aims to prevent both overtreatment and undertreatment in localized prostate cancer, facilitating shared decision making.

Cone beam computed tomography (CBCT) is a widely available modality, but its clinical utility has been limited by low detail conspicuity and quantitative accuracy. Convenient post-reconstruction denoising is subject to back projected patterned residual, but joint denoise-reconstruction is typically computationally expensive and complex. In this study, we develop and evaluate a novel Metric-learning guided wavelet transform reconstruction (MEGATRON) approach to enhance image domain quality with projection-domain processing. Projection domain based processing has the benefit of being simple, efficient, and compatible with various reconstruction toolkit and vendor platforms. However, they also typically show inferior performance in the final reconstructed image, because the denoising goals in projection and image domains do not necessarily align. Motivated by these observations, this work aims to translate the demand for quality enhancement from the quantitative image domain to the more easily operable projection domain. Specifically, the proposed paradigm consists of a metric learning module and a denoising network module. Via metric learning, enhancement objectives on the wavelet encoded sinogram domain data are defined to reflect post-reconstruction image discrepancy. The denoising network maps measured cone-beam projection to its enhanced version, driven by the learnt objective. In doing so, the denoiser operates in the convenient sinogram to sinogram fashion but reflects improvement in reconstructed image as the final goal. Implementation-wise, metric learning was formalized as optimizing the weighted fitting of wavelet subbands, and a res-Unet, which is a Unet structure with residual blocks, was used for denoising. To access quantitative reference, cone-beam projections were simulated using the X-ray based Cancer Imaging Simulation Toolkit (XCIST). In both learning modules, a data set of 123 human thoraxes, which was from Open-Source Imaging Consortium (OSIC) Pulmonary Fibrosis Progression challenge, was used. Reconstructed CBCT thoracic images were compared against ground truth FB and performance was assessed in root mean square error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM). MEGATRON achieved RMSE in HU value, PSNR, and SSIM were 30.97 ± 4.25, 37.45 ± 1.78, and 93.23 ± 1.62, respectively. These values are on par with reported results from sophisticated physics-driven CBCT enhancement, demonstrating promise and utility of the proposed MEGATRON method. We have demonstrated that incorporating the proposed metric learning into sinogram denoising introduces awareness of reconstruction goal and improves final quantitative performance. The proposed approach is compatible with a wide range of denoiser network structures and reconstruction modules, to suit customized need or further improve performance.

BACKGROUND: Artificial intelligence (AI) has transformative potential to support prehospital clinicians, emergency physicians, and trauma surgeons in acute traumatic injury care. This scoping review examines the literature evaluating AI models using prehospital features to support early traumatic injury care. METHODS: We conducted a systematic search in August 2023 of PubMed, Embase, and Web of Science. Two independent reviewers screened titles/abstracts, with a third reviewer for adjudication, followed by a full-text analysis. We included original research and conference presentations evaluating AI models-machine learning (ML), deep learning (DL), and natural language processing (NLP)-that used prehospital features or features available immediately upon emergency department arrival. Review articles were excluded. The same investigators extracted data and systematically categorized outcomes to ensure consistency and transparency. We calculated kappa for interrater reliability and descriptive statistics. RESULTS: We identified 1050 unique publications, with 49 meeting inclusion criteria after title and abstract review (kappa 0.58) and full-text review. Publications increased annually from 2 in 2007 to 10 in 2022. Geographic analysis revealed a 61% focus on data from the United States. Studies were predominantly retrospective (88%), used local (45%) or national level (41%) data, focused on adults only (59%) or did not specify adults or pediatrics (27%), and 57% encompassed both blunt and penetrating injury mechanisms. The majority used machine learning (88%) alone or in conjunction with DL or NLP, and the top three algorithms used were support vector machine, logistic regression, and random forest. The most common study objectives were to predict the need for critical care and life-saving interventions (29%), assist in triage (22%), and predict survival (20%). CONCLUSIONS: A small but growing body of literature described AI models based on prehospital features that may support decisions made by dispatchers, Emergency Medical Services clinicians, and trauma teams in early traumatic injury care.

Band structure and Dirac degeneracy are essential features of sonic crystals/acoustic metamaterials to achieve advanced control of exciting wave effects. In this work, we explore a deep learning approach for the design of phononic crystals with desired dispersion. A plane wave expansion method is utilized to establish the dataset relation between the structural parameters and the energy band features. Subsequently, a multilayer perceptron model trained using the dataset can yield accurate predictions of wave behavior. Based on the trained model, we further impose a re-learning process around a targeted frequency, by which Dirac degeneracy and double Dirac degeneracy can be embedded into the band structures. Our study enables the deep learning approach as a reliable design strategy for Dirac structures/metamaterials, opening up the possibilities for intriguing wave physics associated with Dirac cone. [ABSTRACT FROM AUTHOR]

Non-contact photothermal pump-probe methodologies such as Frequency-Domain Thermo-Reflectance (FDTR) systems facilitate the examination of thermal characteristics spanning semiconductor materials and their associated interfaces. We underscore the significance of meticulous measurements and precise error estimation attained through the analysis of both amplitude and phase data in Thermo-Reflectance (TR). The precision of the analytical estimation hinges greatly on the assumptions made before implementing the method and notably showcases a decrease in errors when both the amplitude and phase are incorporated as input parameters. We demonstrate that frequency-domain calculations can attain high precision in measurements, with error estimations in thermal conductivity (k), thermal boundary resistance (Rth), and thermal diffusivity (α) as low as approximately 2.4%, 2.5%, and 3.0%, respectively. At the outset, we evaluate the uncertainty arising from the existence of local minima when analyzing data acquired via FDTR, wherein both the phase and amplitude are concurrently utilized for the assessment of cross-plane thermal transport properties. Expanding upon data analysis techniques, particularly through advanced deep learning approaches, can significantly enhance the accuracy and precision of predictions when analyzing TR data across a spectrum of modulation frequencies. Deep learning models enhance the quality of fitting and improve the accuracy and precision of uncertainty estimation compared to traditional Monte Carlo simulations. This is achieved by providing suitable initial guesses for data fitting, thereby enhancing the overall performance of the analysis process. [ABSTRACT FROM AUTHOR]

To accelerate the design of a high-power microwave device, a deep learning assisted multi-objective optimization method is used to optimize a Ka-band relativistic backward-wave oscillator (RBWO) operating with a low magnetic field. Particle-in-cell simulation results show that the optimized RBWO with a tooth-shaped slow wave structure (SWS) can generate microwave pulses with an output power of 1.24 GW and an operating frequency of 26.8 GHz under a diode voltage of 623.3 kV, and the diode current is 6.56 kA at a guiding magnetic field of 0.8 T. Compared with the original RBWO, the output power of the optimized RBWO has been increased by 201.2%, and the beam-to-microwave conversion efficiency has increased from 10.0% to 30.3%. The detailed analysis reveals that in an overmoded RBWO with low guiding magnetic fields, the introduction of a tooth-shaped SWS is beneficial to mode competition, improves output power, and decreases microwave starting time. [ABSTRACT FROM AUTHOR]

An accurate estimation of three-dimensional (3D) temperature fields in channel flows is challenging but critical for many important applications such as heat exchangers, radiation energy collectors, and enhanced geothermal systems. In this paper, we demonstrate the possibility of inferring temperature fields from concentration fields for laminar convection flows in a 3D channel using a machine learning (ML) approach. The study involves generation of data using 3D numerical simulations, application of deep learning methodology using conditional generative adversarial networks (cGANs), and analysis of how dataset selection affects model performance. The model is also tested for applicability in different convection scenarios. Results show that cGANs can successfully infer temperature fields from concentration fields, and the reconstruction accuracy is sensitive to the training dataset selected. In this study, we demonstrate how ML can be used to overcome the limitations of traditional heat and mass analogy functions widely used in heat transfer research. [ABSTRACT FROM AUTHOR]

Influencer marketing, in which companies sponsor social media personalities to promote their brands, has exploded in popularity in recent years. One common criterion for selecting an influencer partner is popularity. While some firms collaborate with "mega" influencers with millions of followers, other firms partner with "micro" influencers with only several thousand followers, but who also cost less to sponsor. To quantify this trade-off between popularity and cost, the authors develop a framework for estimating the follower elasticity of impressions (FEI), which measures a video's percentage gain in impressions (i.e., views) corresponding to a percentage increase in the number of followers of its creator. Computing FEI involves estimating the causal effect of an influencer's popularity on the view counts of their videos, which is achieved through a combination of (1) a unique data set collected from TikTok, (2) a representation learning model for quantifying video content, and (3) a machine learning–based causal inference method. The authors find that FEI is always positive, averaging.10, but often nonlinearly related to follower size. They examine the factors that predict variation in these FEI curves and show how firms can use these results to better determine influencer partnerships. [ABSTRACT FROM AUTHOR]

The rapid development of deep learning techniques has driven the emergence of a neural network-based variational Monte Carlo (VMC) method (referred to as FermiNet), which has manifested high accuracy and strong predictive power in the electronic structure calculations of atoms, molecules, and some periodic systems. Recently, the implementation of the effective core potential (ECP) scheme has further facilitated more efficient calculations in practice. However, there is still a lack of comprehensive assessments of the ECP's performance under the FermiNet. In this work, we set sail to fill this gap by conducting extensive tests on the first two row elements regarding their atomic, spectral, and molecular properties. Our major finding is that, in general, the qualities of ECPs have been correctly reflected under FermiNet. Two recently built ECP tables, namely, correlation consistent ECP (ccECP) and energy consistent correlated electron pseudopotential (eCEPP), seem to prevail in terms of overall performance. In particular, ccECP performs slightly better on spectral precision and covers more elements, while eCEPP is more systematically built from both shape and energy consistency and better treats the core polarization. On the other hand, the high accuracy of the all-electron calculations is hindered by the absence of relativistic effects as well as the numerical instabilities in some heavier elements. Finally, with further in-depth discussions, we generate possible directions for developing and improving FermiNet in the near future. [ABSTRACT FROM AUTHOR]

Flat bands in 2D twisted materials are key to the realization of correlation-related exotic phenomena. However, a flat band often was achieved in the large system with a very small twist angle, which enormously increases the computational and experimental complexity. In this work, we proposed group-V twisted bilayer materials, including P, As, and Sb in the β phase with large twist angles. The band structure of twisted bilayer materials up to 2524 atoms has been investigated by a deep learning method DeepH, which significantly reduces the computational time. Our results show that the bandgap and the flat bandwidth of twisted bilayer β-P, β-As, and β-Sb reduce gradually with the decreasing of twist angle, and the ultra-flat band with bandwidth approaching 0 eV is achieved. Interestingly, we found that a twist angle of 9.43° is sufficient to achieve the band flatness for β-As comparable to that of twist bilayer graphene at the magic angle of 1.08°. Moreover, we also find that the bandgap reduces with decreasing interlayer distance while the flat band is still preserved, which suggests interlayer distance as an effective routine to tune the bandgap of flat band systems. Our research provides a feasible platform for exploring physical phenomena related to flat bands in twisted layered 2D materials. [ABSTRACT FROM AUTHOR]

Online ride-hailing services provide additional transportation capability by recruiting private vehicles to meet people's growing travel demand. To ensure the profitability of drivers and platforms, pick-up efficiency and frequency must be maintained at high levels. Therefore, consistency between the spatial distribution of drivers and that of travel demand becomes a key issue to address. This paper proposes a prediction-based iterative Kuhn-Munkres approach for service vehicle reallocation in the context of large-scale online ride-hailing. Firstly, preliminaries are formally defined and a novel mathematical model for the problem is proposed. Secondly, a deep spatio-temporal residual perception network is designed to accurately predict travel demand. Thirdly, an iterative Kuhn-Munkres approach combined with an improved A-Star algorithm is developed to reallocate service vehicles to spatial locations according to their distinct travel demand densities. Finally, extensive experiments are conducted to evaluate and verify the performance of the proposed approach. [ABSTRACT FROM AUTHOR]

The design of heterogeneous catalysts generally involves optimizing the reactivity descriptor of adsorption energy, which is inevitably governed by the structure of surface-active sites. A prerequisite for understanding the structure–properties relationship is the precise identification of real surface-active site structures, rather than relying on conceived structures derived from bulk alloy properties. However, it remains a formidable challenge due to the dynamic nature of nanoalloys during catalytic reactions and the lack of accurate and efficient interatomic potentials for simulations. Herein, a generalizable deep-learning potential for the Ag–Pd–F system is developed based on a dataset encompassing the bulk, surface, nanocluster, amorphous, and point defected configurations with diverse compositions to achieve a comprehensive description of interatomic interactions, facilitating precise prediction of adsorption energy, surface energy, formation energy, and diffusion energy barrier and is utilized to investigate the structural evolutions of AgPd nanoalloys during fluorination. The structural evolutions involve the inward diffusion of F, the outward diffusion of Ag in Ag@Pd nanoalloys, the formation of surface AgFx species in mixed and Janus AgPd nanoalloys, and the shape deformation from cuboctahedron to sphere in Ag and Pd@Ag nanoalloys. Moreover, the effects of atomic diffusion and dislocation formation and migration on the reconstructing pathway of nanoalloys are highlighted. It is demonstrated that the stress relaxation upon F adsorption serves as the intrinsic driving factor governing the surface reconstruction of AgPd nanoalloys. [ABSTRACT FROM AUTHOR]

The present work shows that the free energy landscape associated with alanine dipeptide isomerization can be effectively represented by specific interatomic distances without explicit reference to dihedral angles. Conventionally, two stable states of alanine dipeptide in vacuum, i.e., C7eq (β-sheet structure) and C7ax (left handed α-helix structure), have been primarily characterized using the main chain dihedral angles, φ (C–N–Cα–C) and ψ (N–Cα–C–N). However, our recent deep learning combined with the "Explainable AI" (XAI) framework has shown that the transition state can be adequately captured by a free energy landscape using φ and θ (O–C–N–Cα) [Kikutsuji et al., J. Chem. Phys. 156, 154108 (2022)]. In the perspective of extending these insights to other collective variables, a more detailed characterization of the transition state is required. In this work, we employ interatomic distances and bond angles as input variables for deep learning rather than the conventional and more elaborate dihedral angles. Our approach utilizes deep learning to investigate whether changes in the main chain dihedral angle can be expressed in terms of interatomic distances and bond angles. Furthermore, by incorporating XAI into our predictive analysis, we quantified the importance of each input variable and succeeded in clarifying the specific interatomic distance that affects the transition state. The results indicate that constructing a free energy landscape based on the identified interatomic distance can clearly distinguish between the two stable states and provide a comprehensive explanation for the energy barrier crossing. [ABSTRACT FROM AUTHOR]

Several enhanced sampling techniques rely on the definition of collective variables to effectively explore free energy landscapes. The existing variables that describe the progression along a reactive pathway offer an elegant solution but face a number of limitations. In this paper, we address these challenges by introducing a new path-like collective variable called the "deep-locally non-linear-embedding," which is inspired by principles of the locally linear embedding technique and is trained on a reactive trajectory. The variable mimics the ideal reaction coordinate by automatically generating a non-linear combination of features through a differentiable generalized autoencoder that combines a neural network with a continuous k-nearest neighbor selection. Among the key advantages of this method is its capability to automatically choose the metric for searching neighbors and to learn the path from state A to state B without the need to handpick landmarks a priori. We demonstrate the effectiveness of DeepLNE by showing that the progression along the path variable closely approximates the ideal reaction coordinate in toy models, such as the Müller-Brown potential and alanine dipeptide. Then, we use it in the molecular dynamics simulations of an RNA tetraloop, where we highlight its capability to accelerate transitions and estimate the free energy of folding. [ABSTRACT FROM AUTHOR]

GoogLeNet is a pre-trained Convolutional Neural Network (CNN) that allows transfer learning and has achieved high recognition rates in image classification tasks. A Recurrence Plot (RP) is an imaging method that depicts the recurrence of the state space system using coloured points and lines in 2D images. This work contributes to facilitating time series feature extraction by proposing a method that applies the GoogLeNet to time series images obtained with RP. The developed method is tested using simulated time series and selected time series from the M3 competition dataset. The results shows that the transfer learning approach allowed the extraction of business time series features by means of a GoogLeNet fine-tuned using 100 simulated time series. The combination of GoogLeNet and RPs outperforms the alternative and easier combination of GoogLeNet and plots of the time series and support the convenience of the RP transformation step. This application of deep learning techniques to business time series imaging offers opportunity for further developments. [ABSTRACT FROM AUTHOR]

Metal organic frameworks (MOFs) are crystalline, three-dimensional structures with high surface areas and tunable porosities. Made from metal nodes connected by organic linkers, the exact properties of a given MOF are determined by node and linker choice. MOFs hold promise for numerous applications, including gas capture and storage. M2(4,4′-dioxidobiphenyl-3,3′-dicarboxylate)—henceforth simply M2(dobpdc), with M = Mg, Mn, Fe, Co, Ni, Cu, or Zn—is regarded as one of the most promising structures for CO2 capture applications. Further modification of the MOF with diamines or tetramines can significantly boost gas species selectivity, a necessity for the ultra-dilute CO2 concentrations in the direct-air capture of CO2. There are countless potential diamines and tetramines, paving the way for a vast number of potential sorbents to be probed for CO2 adsorption properties. The number of amines and their configuration in the MOF pore are key drivers of CO2 adsorption capacity and kinetics, and so a validation of computational prediction of these quantities is required to suitably use computational methods in the discovery and screening of amine-functionalized sorbents. In this work, we study the predictive accuracy of density functional theory and related calculations on amine loading and configuration for one diamine and two tetramines. In particular, we explore the Perdew–Burke–Ernzerhof (PBE) functional and its formulation for solids (PBEsol) with and without the Grimme-D2 and Grimme-D3 pairwise corrections (PBE+D2/3 and PBEsol+D2/3), two revised PBE functionals with the Grimme-D2 and Grimme-D3 pairwise corrections (RPBE+D2/3 and revPBE+D2/3), and the nonlocal van der Waals correlation (vdW-DF2) functional. We also investigate a universal graph deep learning interatomic potential's (M3GNet) predictive accuracy for loading and configuration. These results allow us to identify a useful screening procedure for configuration prediction that has a coarse component for quick evaluation and a higher accuracy component for detailed analysis. Our general observation is that the neural network-based potential can be used as a high-level and rapid screening tool, whereas PBEsol+D3 gives a completely qualitatively predictive picture across all systems studied, and can thus be used for high accuracy motif predictions. We close by briefly exploring the predictions of relative thermal stability for the different functionals and dispersion corrections. [ABSTRACT FROM AUTHOR]

The current focus in Autonomous Experimentation (AE) is on developing robust workflows to conduct the AE effectively. This entails the need for well-defined approaches to guide the AE process, including strategies for hyperparameter tuning and high-level human interventions within the workflow loop. This paper presents a comprehensive analysis of the influence of initial experimental conditions and in-loop interventions on the learning dynamics of Deep Kernel Learning (DKL) within the realm of AE in scanning probe microscopy. We explore the concept of the "seed effect," where the initial experiment setup has a substantial impact on the subsequent learning trajectory. Additionally, we introduce an approach of the seed point interventions in AE allowing the operator to influence the exploration process. Using a dataset from Piezoresponse Force Microscopy on PbTiO3 thin films, we illustrate the impact of the "seed effect" and in-loop seed interventions on the effectiveness of DKL in predicting material properties. The study highlights the importance of initial choices and adaptive interventions in optimizing learning rates and enhancing the efficiency of automated material characterization. This work offers valuable insights into designing more robust and effective AE workflows in microscopy with potential applications across various characterization techniques. [ABSTRACT FROM AUTHOR]

In nowadays production, fluctuations in demand, shortening product life-cycles, and highly configurable products require an adaptive and robust control approach to maintain competitiveness. This approach must not only optimise desired production objectives but also cope with unforeseen machine failures, rush orders, and changes in short-term demand. Previous control approaches were often implemented using a single operations layer and a standalone deep learning approach, which may not adequately address the complex organisational demands of modern manufacturing systems. To address this challenge, we propose a hyper-heuristics control model within a semi-heterarchical production system, in which multiple manufacturing and distribution agents are spread across pre-defined modules. The agents employ a deep reinforcement learning algorithm to learn a policy for selecting low-level heuristics in a situation-specific manner, thereby leveraging system performance and adaptability. We tested our approach in simulation and transferred it to a hybrid production environment. By that, we were able to demonstrate its multi-objective optimisation capabilities compared to conventional approaches in terms of mean throughput time, tardiness, and processing of prioritised orders in a multi-layered production system. The modular design is promising in reducing the overall system complexity and facilitates a quick and seamless integration into other scenarios. [ABSTRACT FROM AUTHOR]