Complex Input Data Research Articles

Artificial intelligence and classification of mature lymphoid neoplasms.

Hematologists, geneticists, and clinicians came to a multidisciplinary agreement on the classification of lymphoid neoplasms that combines clinical features, histological characteristics, immunophenotype, and molecular pathology analyses. The current classification includes the World Health Organization (WHO) Classification of tumours of haematopoietic and lymphoid tissues revised 4th edition, the International Consensus Classification (ICC) of mature lymphoid neoplasms (report from the Clinical Advisory Committee 2022), and the 5th edition of the proposed WHO Classification of haematolymphoid tumours (lymphoid neoplasms, WHO-HAEM5). This article revises the recent advances in the classification of mature lymphoid neoplasms. Artificial intelligence (AI) has advanced rapidly recently, and its role in medicine is becoming more important as AI integrates computer science and datasets to make predictions or classifications based on complex input data. Summarizing previous research, it is described how several machine learning and neural networks can predict the prognosis of the patients, and classified mature B-cell neoplasms. In addition, new analysis predicted lymphoma subtypes using cell-of-origin markers that hematopathologists use in the clinical routine, including CD3, CD5, CD19, CD79A, MS4A1 (CD20), MME (CD10), BCL6, IRF4 (MUM-1), BCL2, SOX11, MNDA, and FCRL4 (IRTA1). In conclusion, although most categories are similar in both classifications, there are also conceptual differences and differences in the diagnostic criteria for some diseases. It is expected that AI will be incorporated into the lymphoma classification as another bioinformatics tool.

A review of mechanistic learning in mathematical oncology.

Mechanistic learning refers to the synergistic combination of mechanistic mathematical modeling and data-driven machine or deep learning. This emerging field finds increasing applications in (mathematical) oncology. This review aims to capture the current state of the field and provides a perspective on how mechanistic learning may progress in the oncology domain. We highlight the synergistic potential of mechanistic learning and point out similarities and differences between purely data-driven and mechanistic approaches concerning model complexity, data requirements, outputs generated, and interpretability of the algorithms and their results. Four categories of mechanistic learning (sequential, parallel, extrinsic, intrinsic) of mechanistic learning are presented with specific examples. We discuss a range of techniques including physics-informed neural networks, surrogate model learning, and digital twins. Example applications address complex problems predominantly from the domain of oncology research such as longitudinal tumor response predictions or time-to-event modeling. As the field of mechanistic learning advances, we aim for this review and proposed categorization framework to foster additional collaboration between the data- and knowledge-driven modeling fields. Further collaboration will help address difficult issues in oncology such as limited data availability, requirements of model transparency, and complex input data which are embraced in a mechanistic learning framework.

Development of experimentally-validated models for nocturnal cooling of thermal systems with photovoltaic thermal (PVT) modules

This present work describes the development of an experimentally-validated (i) detailed mathematical model and (ii) simplified regression model, for the purpose of investigating the thermal performance of photovoltaic thermal (PVT) modules undergoing nocturnal cooling of thermal systems. The detailed mathematical model is useful for hourly, detailed simulation of thermal systems, while the simplified regression model is applicable to situations needing rapid thermal performance assessment and design of PVT systems for cooling. Such models currently do not exist, owing to the difficulty in quantifying relevant thermal radiation parameters. Thus, another outcome of this work is the foundation of a test method for quantifying cooling performance of PVT systems. In developing the detailed hourly model, the mean absolute error between the detailed model and the experiment was found to be 0.74 ± 0.91 K for predicting the fluid outlet temperature. A parametric analysis was conducted to study the effect of some physical and environmental parameters on the PVT performance, and results showed that the nocturnal cooling power increased by up to 45 W/m2 under favorable radiative cooling conditions. The simplified model underpredicts the outlet fluid temperature by up to 2 °C, especially in winter, but serves as a quick, practical design tool, eliminating the necessity of complex input data and heavy computational demands. The simplified PVT performance curve has versatile applications to various thermal energy systems that benefit from integrating PVT nocturnal cooling, such as underground thermal storage reservoirs utilized with ground source heat pumps, buildings, swimming pools, and phase change material applications.

Flexible Quality Control for Protein Turnover Rates Using d2ome.

Bioinformatics tools are used to estimate in vivo protein turnover rates from the LC-MS data of heavy water labeled samples in high throughput. The quantification includes peak detection and integration in the LC-MS domain of complex input data of the mammalian proteome, which requires the integration of results from different experiments. The existing software tools for the estimation of turnover rate use predefined, built-in, stringent filtering criteria to select well-fitted peptides and determine turnover rates for proteins. The flexible control of filtering and quality measures will help to reduce the effects of fluctuations and interferences to the signals from target peptides while retaining an adequate number of peptides. This work describes an approach for flexible error control and filtering measures implemented in the computational tool d2ome for automating protein turnover rates. The error control measures (based on spectral properties and signal features) reduced the standard deviation and tightened the confidence intervals of the estimated turnover rates.

Automated Mapping of Wetland Ecosystems: A Study Using Google Earth Engine and Machine Learning for Lotus Mapping in Central Vietnam

Wetlands are highly productive ecosystems with the capability of carbon sequestration, providing an effective solution for climate change. Recent advancements in remote sensing have improved the accuracy in the mapping of wetland types, but there remain challenges in accurate and automatic wetland mapping, with additional requirements for complex input data for a number of wetland types in natural habitats. Here, we propose a remote sensing approach using the Google Earth Engine (GEE) to automate the extraction of water bodies and mapping of growing lotus, a wetland type with high economic and cultural values in central Vietnam. Sentinel-1 was used for water extraction with the K-Means clustering, whilst Sentinel-2 was combined with the machine learning smile Random Forest (sRF) and smile Gradient Tree Boosting (sGTB) models to map areas with growing lotus. The water map was derived from S-1 images with high confidence (F1 = 0.97 and Kappa coefficient = 0.94). sGTB outperformed the sRF model to deliver a growth map with a high accuracy (overall accuracy = 0.95, Kappa coefficient = 0.92, Precision = 0.93, and F1 = 0.93). The total lotus area was estimated at 145 ha and was distributed in the low land of the study site. Our proposed framework is a simple and reliable mapping technique, has a scalable potential with the GEE, and is capable of extension to other wetland types for large-scale mapping worldwide.

“Symcrete” memory Model with Lazy Initialization and Objects of Symbolic Sizes in KLEE

Dynamic symbolic execution is a well-known technique for testing applications. It introduces symbolic variables – program data with no concrete value at the moment of instantiation – and uses them to systematically explore the execution paths in a program under analysis. However, not every value can be easily modelled as symbolic: for instance, some values may take values from restricted domains or have complex invariants, hard enough to model using existing logic theories, despite it is not a problem for concrete computations. In this paper, we propose an implementation of infrastructure for dealing with such “hard-to-be-modelled” values. We take the approach known as symcrete execution and implement its robust and scalable version in the well-known KLEE symbolic execution engine. We use this infrastructure to support the symbolic execution of LLVM programs with complex input data structures and input buffers with indeterminate sizes.

Passenger Flow Prediction of Scenic Spots in Jilin Province Based on Convolutional Neural Network and Improved Quantile Regression Long Short-Term Memory Network

Passenger flow is an important benchmark for measuring tourism benefits, and accurate tourism passenger flow prediction is of great significance to the government and related tourism enterprises and can promote the sustainable development of China’s tourism industry. For daily passenger flow time series data, a passenger flow forecasting method based on convolutional neural network (CNN) and improved quantile regression long short-term memory network (QRLSTM), denoted as CNN-IQRLSTM, is proposed with reconstructed correlation features and in the form of sliding windows as inputs. First, four discrete variables such as whether the day is a weekend and holiday are created by time; then, a sliding window of width 42 is used to pass the passenger flow data into the network sequentially; finally, the loss function of the sparse Laplacian improved QRLSTM is introduced for passenger flow prediction, and the point prediction and interval prediction results under different quartiles are obtained. The application of quantile regression captures the overall picture of the data, enhances the robustness, fit, predictive power and nonlinear processing capability of neural networks, and fills the gap between quantile regression and neural network methods in the field of passenger flow prediction. CNN can effectively handle complex input data, and the improved nonlinear QR model can provide passenger flow quantile prediction information. The method is applied to the tourism traffic prediction of four 5A scenic spots in Jilin Province, and the effectiveness of the method is verified. The results show that the method proposed in this paper fits best in point prediction and has higher prediction accuracy. The MAPE of the Changbai Mountain dataset was 0.07, the MAPE of the puppet palace museum dataset was 0.05, the fit of the Sculpture Park dataset reached 93%, and the fit of the net moon lake dataset was as high as 99%. Meanwhile, the interval prediction results show that the method has a larger interval coverage as well as a smaller interval average width, which improves the prediction efficiency. In 95% of the interval predictions, the interval coverage of Changbai Mountain data is 99% and the interval average width is 0.49. It is a good reference value for the management of different scenic spots.

A high-resolution computationally-efficient spatiotemporal model for estimating daily PM2.5 concentrations in Beijing, China

Exposure to ambient air pollution is the largest environmental source of global disease burden. Accurate assessment of air pollutant exposure at finely-resolved spatial and temporal scales is critical for valid estimation of health effects. Limitations of exposure assessment approaches include computational burden and unreliable or complex input data. We aimed to develop an accurate, easy-to-use, high-resolution statistical model to estimate ambient PM2.5 concentrations in Beijing. We implemented a model that estimates long- and short-term trends in pollutant concentrations based on observations from regulatory monitors, geographic covariates, and spatial smoothing. It also allows for inclusion of spatiotemporal covariates. We used observations from 19 monitors around Beijing and 90 geographic covariates, including road density, meteorological covariates, population and building density, land use types, topography, and vegetation cover to produce predictions of daily PM2.5 concentrations for 2015–2017. The model yields predictions at any geographic point, and here we summarize results for a 500m × 500m scale. The daily and long-term average cross validated R2 were 0.96 and 0.93, with RMSE of 13.1 and 1.7, respectively. The addition of temperature as a spatiotemporal covariate did not change the results materially. Over the study period, annual and seasonal PM2.5 concentrations in Beijing declined substantially, although they remained high relative to recommended level by WHO. We developed an easy-to-use, less computationally-intensive statistical model utilizing readily available input data to provide highly accurate estimates of PM2.5 concentrations in Beijing at a fine spatiotemporal scale. Model predictions have already been used in a publicly available mobile application designed to provide exposure estimates for epidemiological analyses of the effects of PM2.5 among residents of Beijing.

Prognostic Significance of Simple Scoring Systems in the Prediction of Diffuse Peritonitis Morbidity and Mortality

Introduction: Diffuse peritonitis is a serious disease. It is often addressed within urgent management of an unstable patient in shock. The therapy consists of treatment of the source of peritonitis, decontamination of the abdominal cavity, stabilization of the patient and comprehensive resuscitation care in an intensive care unit. A number of scoring systems to determine patient prognosis are available, but most of them require complex input data, making their practical application a substantial problem. Objective: Our aim was to assess simple scoring systems within a cohort, evaluate the level of mortality, morbidity, and duration of hospital stay, followed by a comparison of the acquired data with the literature and determination of an easily implementable scoring system for use in clinical practice. Material and Methods: We evaluated a group of patients with diffuse peritonitis who underwent surgery in the 2015–2019 period. Medical history, surgical findings, and paraclinical examinations were used as the input for four scoring systems commonly used in practice—MPI, qSOFA, ECOG, and ASA. We compared the results between the systems and with the literature. Results: Our cohort included 274 patients diagnosed with diffuse peritonitis. Mortality was 22.6%, morbidity 73.4%, with a 25.2 day average duration of hospital stay. Mortality and morbidity increased with rising MPI and qSOFA, well-established scoring systems, but also with rising ASA and ECOG, similarly to MPI and qSOFA. Conclusions: The utilized scoring systems correlated well with the severity of the condition and with predicted mortality and morbidity as reported in the literature. Simple scoring systems primarily used in other indications (i.e., ASA and ECOG) have a similar predictive value in our cohort as commonly used systems (MPI, qSOFA). We recommend them in routine clinical practice due to their simplicity.

Автоматическое тестирование LLVM-программ со сложными входными структурами данных

Symbolic execution is a widely used approach for automatic regression test generation and bug and vulnerability finding. The main goal of this paper is to present a practical symbolic execution-based approach for LLVM programs with complex input data structures. The approach is based on the well-known idea of lazy initialization, which frees the user from providing constraints on input data structures manually. Thus, it provides us with a fully automatic symbolic execution of even complex program. Two lazy initialization improvements are proposed for segmented memory models: one based on timestamps and one based on type information. The approach is implemented in the KLEE symbolic virtual machine for the LLVM platform and tested on real C data structures — lists, binomial heaps, AVL trees, red-black trees, binary trees, and tries.

Improving Accuracy of Tomato Plant Disease Diagnosis Based on Deep Learning With Explicit Control of Hidden Classes.

Recognizing plant diseases is a major challenge in agriculture, and recent works based on deep learning have shown high efficiency in addressing problems directly related to this area. Nonetheless, weak performance has been observed when a model trained on a particular dataset is evaluated in new greenhouse environments. Therefore, in this work, we take a step towards these issues and present a strategy to improve model accuracy by applying techniques that can help refine the model’s generalization capability to deal with complex changes in new greenhouse environments. We propose a paradigm called “control to target classes.” The core of our approach is to train and validate a deep learning-based detector using target and control classes on images collected in various greenhouses. Then, we apply the generated features for testing the inference of the system on data from new greenhouse conditions where the goal is to detect target classes exclusively. Therefore, by having explicit control over inter- and intra-class variations, our model can distinguish data variations that make the system more robust when applied to new scenarios. Experiments demonstrate the effectiveness and efficiency of the proposed approach on our extended tomato plant diseases dataset with 14 classes, from which 5 are target classes and the rest are control classes. Our detector achieves a recognition rate of target classes of 93.37% mean average precision on the inference dataset. Finally, we believe that our study offers valuable guidelines for researchers working in plant disease recognition with complex input data.

Four Aspects Affecting Health Economic Decision Models and Their Validation.

Health care decision makers in many jurisdictions use cost-effectiveness analysis based on health economic decision models for policy decisions regarding coverage and price negotiation for medicines and medical devices. While validation of health economic decision models has always been considered important, many reviews of model-based cost-effectiveness studies report limitations regarding their validation. The current opinion paper discusses four aspects of current health economic decision modeling with relevance for future directions in model validation: increased use of complex models, international cooperation, open-source modeling, and stakeholder involvement. First, new, more complex clinical study designs and treatment strategies may require relatively complex model structures and/or input data analyses. Simultaneously, more widespread technical knowledge along with wider data availability have led to a broader range of model types. This puts extra requirements on model validation and transparency. Second, increased international cooperation of policy makers and, in particular, health technology assessment (HTA) authorities in performing model assessments is discussed in relation to the repeated use of health economic models (multi-use disease models). We argue such coordinated efforts may benefit model validity. Third, open-source modeling is discussed as one possible answer to increased transparency requirements. Finally, involvement of all relevant stakeholders throughout the whole decision process is an ongoing development that necessarily also includes health economic modeling. We argue this implies that model validity should be considered in a broader perspective, with more focus on conceptual modeling, model transparency, accuracy requirements, and choice of relevant model outcomes than previously.

Reservoir inflow predicting model based on machine learning algorithm via multi‐model fusion: A case study of Jinshuitan river basin

Flood prevention and disaster mitigation have a great impact on people's lives and properties, and so it is urgent to realise high-accuracy inflow predictions for flood early warning. To this end, a prediction model based on a machine learning algorithm via a multimodel combination method is proposed to predict the inflow of Jinshuitan reservoir. Firstly, a data formatting scheme called the ’hydrological regime profile‘ is designed for input data. The whole data set is partitioned into a low-flow subset and a high-flow subset. Considering the high dimensions of the complex input data, convolutional neural networks (CNN), EXtreme gradient Boosting model (XGBoost) and a partial least squares model (PLS) are used. In the CNN and XGBoost models, a special loss function weighted on inflow is designed to improve the performance on high-inflow predictions. Finally, a multi-model combination method is proposed to improve the prediction performance. Compared with XGBoost, CNN and PLS, the root mean square error of the combined model is reduced by 41.64%, 72.29% and 3.41%, respectively. As a consequence, the combined model is able to predict the inflows with higher accuracy compared to the single models.

A Survey on Pipelined FFT Hardware Architectures

The field of pipelined FFT hardware architectures has been studied during the last 50 years. This paper is a survey that includes the main advances in the field related to architectures for complex input data and power-of-two FFT sizes. Furthermore, the paper is intended to be educational, so that the reader can learn how the architectures work. Finally, the paper divides the architectures into serial and parallel. This classification puts together those architectures that are conceived for a similar purpose and, therefore, are comparable.

Design and Analysis of the Fractional-Order Complex Least Mean Square (FoCLMS) Algorithm

In this work, a new class of stochastic gradient algorithm is developed based on fractional calculus. Unlike the existing algorithms, the concept of complex fractional gradient is introduced by employing Caputo’s fractional derivative which results in a fractional steepest descent algorithm and a fractional-order complex LMS (FoCLMS) algorithm. We demonstrate that with the Caputo’s fractional gradient definition, the Weiner solution remains invariant. Convergence analysis of the proposed FoCLMS algorithm is presented for both transient and steady state scenarios. Consequently, expressions for the learning curves and steady state EMSE are derived. Our theoretical developments are validated by simulation experiments. Extensive simulations are presented to investigate all possible scenarios: channel with negative weights and real input data, channel with positive weights and complex input data, and channel with complex weights and complex input data.

Reinforcement learning in surgery

Patients and physicians make essential decisions regarding diagnostic and therapeutic interventions. These actions should be performed or deferred under time constraints and uncertainty regarding patients’ diagnoses and predicted response to treatment. This may lead to cognitive and judgment errors. Reinforcement learning is a subfield of machine learning that identifies a sequence of actions to increase the probability of achieving a predetermined goal. Reinforcement learning has the potential to assist in surgical decision making by recommending actions at predefined intervals and its ability to utilize complex input data, including text, image, and temporal data, in the decision-making process. The algorithm mimics a human trial-and-error learning process to calculate optimum recommendation policies. The article provides insight regarding challenges in the development and application of reinforcement learning in the medical field, with an emphasis on surgical decision making. The review focuses on challenges in formulating reward function describing the ultimate goal and determination of patient states derived from electronic health records, along with the lack of resources to simulate the potential benefits of suggested actions in response to changing physiological states during and after surgery. Although clinical implementation would require secure, interoperable, livestreaming electronic health record data for use by virtual model, development and validation of personalized reinforcement learning models in surgery can contribute to improving care by helping patients and clinicians make better decisions.

Physical model inversion of the green spectral region to track assimilation rate in almond trees with an airborne nano-hyperspectral imager

Significant advances toward the remote sensing of photosynthetic activity have been achieved in the last decades, including sensor design and radiative transfer model (RTM) development. Nevertheless, finding methods to accurately quantify carbon assimilation across species and spatial scales remains a challenge. Most methods are either empirical and not transferable across scales or can only be applied if highly complex input data are available. Under stress, the photosynthetic rate is limited by the maximum carboxylation rate (Vcmax), which is determined by the leaf biochemistry and the environmental conditions. Vcmax has been connected to plant photoprotective mechanisms, photosynthetic activity and chlorophyll fluorescence emission. Recent RTM developments such as the Soil-Canopy Observation of Photosynthesis and Energy fluxes (SCOPE) model allow the simulation of the sun-induced chlorophyll fluorescence (SIF) and Vcmax effects on the canopy spectrum. This development provides an approach to retrieve Vcmax through RTM model inversion and track assimilation rate. In this study we explore SIF, narrow-band indices and RTM inversion to track changes in photosynthetic efficiency as a function of vegetation stress. We use hyperspectral imagery acquired over an almond orchard under different management strategies which affected the assimilation rates measured in the field. Vcmax used as an indicator of assimilation was retrieved through SCOPE model inversion from pure-tree crown hyperspectral data. The relationships between field-measured assimilation rates and Vcmax retrieved from model inversion were higher (r2 = 0.7–0.8) than when SIF was used alone (r2 = 0.5–0.6) or when traditional vegetation indices were used (r2 = 0.3–0.5). The method was proved successful when applied to two independent datasets acquired at two different dates throughout the season, ensuring its robustness and transferability. When applied to both dates simultaneously, the results showed a unique significant trend between the assimilation measured in the field and Vcmax derived using SCOPE (r2 = 0.56, p < 0.001). This work demonstrates that tracking assimilation in almond trees is feasible using hyperspectral imagery linked to radiative transfer-photosynthesis models.

Improved convolutional neural network‐based quantile regression for regional photovoltaic generation probabilistic forecast

Nowadays, an increasing number of photovoltaic (PV) plants are becoming integrated into one regional power grid. Under this circumstance, the probabilistic forecast of regional PV power generation is of significance for the regional power system operation and control. This study presents a novel probabilistic forecast method for regional PV generation that integrates the convolutional neural network (CNN) with non-linear quantile regression (QR). In this method, the CNN structure is enhanced to extract the non-linear features of the input data and generate the non-linear QR function. As a result, the improved CNN can effectively process high-dimensional and complex input data and the non-linear QR model can provide quantile forecast information of regional PV power. The validity of the proposed method is verified by using it to forecast the regional PV generation from the clustered PV plants in the Weifang region of China.

Gradient analysis of Markov-type control schemes and its applications

This paper presents formulas for gradients of the Average Run Length (ARL) and other Run Length characteristics by control scheme parameters, in the context of Markov Chain approach to analysis of Markov-type control schemes. We illustrate use of these formulas in several problems related to control charting, including (a) design of control schemes, (b) inference on Run Length characteristics based on Phase-1 data and (c) performance analysis for highly complex input data distributions.

Using machine-learning methods to analyze economic loss function of quality management processes

During analysis of quality management systems, their economic component is often analyzed insufficiently. To overcome this issue, it is necessary to withdraw the concept of economic loss functions from tolerance thinking and address it. Input data about economic losses in processes have a complex form, thus, using standard tools to solve this problem is complicated. Use of machine learning techniques allows one to obtain precise models of the economic loss function based on even the most complex input data. Results of such analysis contain data about the true efficiency of a process and can be used to make investment decisions.