Prediction Error Of Model Research Articles

Gear fault diagnosis based on bidimensional time-frequency information theoretic features and error-correcting output codes: A multi-class support vector machine

Fault diagnosis of gears plays an important role in reducing downtime and maximizing efficiency of rotating machinery. Vibration is popular parameter for gear fault detection. The occurrence of faults produces recurring transient impulses in the gear vibration signals. However, these transient features are heavily masked by background noises making it difficult to investigate gear faults. Furthermore, the development of automated fault diagnosis techniques requiring minimal human supervision poses another big challenge. Consequently, in this paper, an automated fault diagnosis technique based on a novel information theoretic (IT)-derived-feature set and an artificial intelligence technique called as error-correcting output codes-support vector machine (ECOC-SVM) is proposed. The gear vibration signals are first processed by continuous wavelet transform to obtain the corresponding time-frequency distributions (TFDs). The TFDs of the faulty signals are then discriminated from those of the healthy ones by introducing IT measures, namely, Kulback-Leibller divergence (KLD), Jensen-Shannon divergence (JSD), Jensen-Rényi divergence (JRD), and Jensen-Tsallis divergence (JTD). These uni-dimensional-IT measures are modified to accommodate the bidimensional TFDs, and the resultant features are referred to as bidimensional time-frequency information theoretic divergence (BTF-ITD) features. The BTF-ITD features are then used to train the ECOC-SVM model. Finally, the trained ECOC-SVM model is used for testing the gear faults. The ECOC approach rectifies the biases and errors in SVM model predictions. The experimental results confirm that the proposed approach provides higher classification accuracy than time-domain features; voting-based-multiclass SVM; and deep learning techniques, such as those based on the stacked sparse autoencoder (SSAE), deep neural network (DNN), and convolution neural network (CNN).

Adopting combination machine learning models could reduce hospital length of stay for oncology patients.

11084 Background: Integra Connect previously created in-patient (IP) admission prediction model based on OCM data. One of the challenges for practices in a value-based care (VBC) program is to provide continuous care-coordination during and after an IP admission. Our objective is to show how adopting combination machine learning (ML) models can predict IP length of stay (LOS). Potential benefits include an overall reduction in LOS which could minimize the risk of hospital acquired conditions. Methods: ML models were trained on 5 major cancer types (Lung Cancer, Multiple Myeloma, Lymphoma, Small Intestine / Colorectal Cancer, High-risk Breast Cancer') from 7 OCM practices of PP4-PP9 data, excluding surgery IP admission. The top 5 cancer types accounted for ~50% of total in-patient admissions and ~50% of total LOS in days. The IP admissions were divided into 4 major cohorts in terms of LOS in days (1-3: class 1, 4-8: class 2, 9-15: class 3, and more than 15: class 4). To reduce the overall LOS, we adopted two ML models: 1) To classify LOS, a multi-classification model (Model1, an eXtreme Gradient Boosted Trees Classifier) and 2) To predict the LOS for classes 1-3, a regression-based model (Model2, an eXtreme Gradient Boosted Trees Regressor with Early Stopping). Class 4 cohort had a wide range (16- 90+ days) for LOS and therefore was excluded from Model2. The other three cohorts belonged to Q1, IQR, and Q4, respectively. Model1 was trained on 4,280 randomly selected sample IP admissions to balance each class and Model2 was trained on 19,636 IP admissions, both with 102 features. Results: The models were tested on PP10-PP11 claims excluding 0 days LOS. Model1 predicted 296 (3,945) IP admissions in class 4 of which 88 were true positives. Assuming at least a 10% reduction in associated total LOS translated to lower LOS of 1-3% for 5 (7) practices. The remaining 3,649 IP admissions were used to predict LOS by Model2. The difference between actual LOS and Predicted LOS was found as 18%-24%. To address model prediction errors, we defined the target LOS to be predicted LOS with an upper bound of at least 10%. This led to reductions of 6% -16% across practices. Finally, combining the results of both ML models we determined that the potential to lower LOS was at least 6% and at most 19%. Conclusions: Our ML models identified opportunities to reduce LOS across multiple OCM cancer practices for 5 major cancer types. We also identified a cohort of patients with critical condition (class 4) that is vital for practice transformation initiatives. These identified reduction in LOS for oncology patients provides cost reduction and quality improvement opportunities in VBC programs. In the future, more opportunities to lower LOS will be explored with a re-admission prediction model.

A prediction method for salt frost resistance of dam rock foundation based on improved BP neural network

Hydraulic resources can be utilized in a cascade manner, with the advantages of energy conservation, economy, and environmental protection. In order to study the anti salt freezing performance of rock foundation in hydroelectric power dams and accurately predict the degree of rock and soil freeze-thaw damage under salt freezing conditions, this paper proposes a prediction method for the anti salt freezing performance of dam rock foundation based on an improved BP neural network. Firstly, freeze-thaw cycle tests were conducted on the rock foundation materials of dams poured with different proportions of fiber reinforced concrete to study the changes in soil mass loss rate and dynamic elastic modulus; Then, based on BP neural network and particle swarm optimization algorithm, a prediction model for rock and soil freeze-thaw damage is established; Finally, using the historical data of the Baihetan hydropower station, optimization was carried out and model prediction errors were compared and analyzed. The method proposed in this article has been verified to have good accuracy and stability, providing guidance for the operation and construction design of hydropower projects.

An Improved Back Propagation Neural Network Based on Differential Evolution and Grey Wolf Optimizer and Its Application in the Height Prediction of Water-Conducting Fracture Zone

Given that the conventional back propagation neural network (BPNN) easily falls into the local optimal solutions, resulting in poor prediction accuracy, an improved BPNN based on the differential evolution and grey wolf optimizer (DEGWO) is proposed, the so-called DEGWO-BPNN. The prediction of the water-conducting fracture zone (WCFZ) height is significant for mine safety operations. A total of 104 sample data are trained and 25 sample data are tested to identify the optimal prediction model. Five evaluation indexes are selected to assess the prediction performance of the models quantitatively. Finally, the DEGWO-BPNN model is applied to a specific engineering case. The main conclusions are as follows: (1) Mining height, mining depth, coal seam dip, panel width, and ratio of hard rock as the main factors affecting the WCFZ height are selected. The topology structure of the model is defined as ‘5-12-1’; (2) the bias between the predicted value and the actual value of the training samples is smaller with an average error of 2.39. Test samples further validate the prediction precision through evaluation indexes. The values of MAE, RMSE, MAPE, and R2 are 2.3952, 3.4674, 5.3148%, and 0.99077, respectively. The prediction accuracy is 94.6852%; (3) ‘Mining Code’, MLR, BPNN, and GWO-BPNN models are treated as the comparison groups. The comparative analysis shows that the prediction performance of ‘Mining Code’ is the worst, while that of DEGWO-BPNN is the best, and it outperforms other algorithms and statistical approaches; (4) the prediction of WCFZ height in the 11601 panel is in line with the actual value. The prediction error of the DEGWO-BPNN model is lower than that of the comparison models. As such, the DEGWO-BPNN model can be well applied to the prediction of WCFZ height and is suitable for coal mines with different regional geological conditions. It can provide a valuable reference for mine safety operations.

A hardness-based constitutive model for numerical simulation of blind rivet with gradient hardness distribution

A286 superalloy blind rivets find extensive applications in the structural connections of aerospace equipment owing to their exceptional strength and heat resistance. However, the gradient hardness distribution within the rivet sleeves complicates achieving precise numerical simulation results with traditional constitutive models. This study aims to establish a constitutive model suitable for materials exhibiting gradient hardness to support the development of A286 superalloy blind rivets. Initially, quasi-static compression experiments were conducted on A286 superalloy samples at different hardness levels (160–324 HV), strain rates (0.01–10 s⁻¹), and temperatures (25–600°C). Subsequently, the influence of material microstructural evolution on flow stress during compression was investigated using electron backscatter diffraction technique. Furthermore, a hardness-based Johnson-Cook (J-C) constitutive model was developed by introducing hardness and offset compensation onto the original J-C model and compared with its counterpart. Finally, the hardness-based model was employed to simulate the bulge compression forming of A286 blind rivets and validated through experimental trials. The findings reveal that dislocation strengthening predominantly contributes to the increased material flow stress with escalating strain rates and sample hardness. Compared to the original J-C model, the stress prediction error of the hardness-based model reduced from 26.8 % to 2.9 %, effectively depicting the variation in material flow stress under different hardness values. The simulated values of bulge morphology, dimensions, and load closely align with experimental data, affirming the efficacy of the hardness-based model in numerically simulating blind rivets with gradient hardness distribution.

Tackling Data Scarcity Challenge through Active Learning in Materials Processing with Electrospray

Machine learning (ML) has been harnessed as a promising modelling tool for materials research. However, small data, or data scarcity, is a bottleneck when incorporating ML in studies involving experimentation. Current experiment planning methods show several disadvantages: one‐factor‐at‐a‐time (OFAT) experimentation became impractical due to limited laboratory resources; conventional design of experiments (DoE) failed to incorporate high‐dimensional features in ML; Surrogate‐based or Bayesian optimization (BO) shifted the goal to optimize material properties rather than guiding training data accumulation. The present research proposes leveraging active learning (AL) to strategically select critical data for experimentation. Two AL strategies, query‐by‐Committee (QBC) algorithm and Greedy method, are benchmarked against random query baseline on various materials datasets. AL is shown to efficiently reduce model prediction errors with minimal additional experiment data. Investigation of hyperparameters revealed benefits of applying AL at an early stage of experimental dataset construction. Moreover, AL is implemented and validated for an in‐house materials development task ‐ electrospray modelling. AL exploration as a paradigm is highlighted to guide experiment design for efficient data accumulation purposes, and its potential for further ML modelling. In doing so, the power of ML is expected to be fully unleashed to experimental researchers.

Research on CC-SSBLS Model-Based Air Quality Index Prediction

Establishing reliable and effective prediction models is a major research priority for air quality parameter monitoring and prediction and is utilized extensively in numerous fields. The sample dataset of air quality metrics often established has missing data and outliers because of certain uncontrollable causes. A broad learning system based on a semi-supervised mechanism is built to address some of the dataset’s data-missing issues, hence reducing the air quality model prediction error. Several air parameter sample datasets in the experiment were discovered to have outlier issues, and the anomalous data directly impact the prediction model’s stability and accuracy. Furthermore, the correlation entropy criteria perform better when handling the sample data’s outliers. Therefore, the prediction model in this paper consists of a semi-supervised broad learning system based on the correlation entropy criterion (CC-SSBLS). This technique effectively solves the issue of unstable and inaccurate prediction results due to anomalies in the data by substituting the correlation entropy criterion for the mean square error criterion in the BLS algorithm. Experiments on the CC-SSBLS algorithm and comparative studies with models like Random Forest (RF), Support Vector Regression (V-SVR), BLS, SSBLS, and Categorical and Regression Tree-based Broad Learning System (CART-BLS) were conducted using sample datasets of air parameters in various regions. In this paper, the root mean square error (RMSE) and mean absolute percentage error (MAPE) are used to judge the advantages and disadvantages of the proposed model. Through the experimental analysis, RMSE and MAPE reached 8.68 μg·m−3 and 0.24% in the Nanjing dataset. It is possible to conclude that the CC-SSBLS algorithm has superior stability and prediction accuracy based on the experimental results.

Short-Term Electrical Load Forecasting Using an Enhanced Extreme Learning Machine Based on the Improved Dwarf Mongoose Optimization Algorithm

Accurate short-term electrical load forecasting is crucial for the stable operation of power systems. Given the nonlinear, periodic, and rapidly changing characteristics of short-term power load forecasts, this paper introduces a novel forecasting method employing an Extreme Learning Machine (ELM) enhanced by an improved Dwarf Mongoose Optimization Algorithm (Local escape Dwarf Mongoose Optimization Algorithm, LDMOA). This method addresses the significant prediction errors of conventional ELM models and enhances prediction accuracy. The enhancements to the Dwarf Mongoose Optimization Algorithm include three key modifications: initially, a dynamic backward learning strategy is integrated at the early stages of the algorithm to augment its global search capabilities. Subsequently, a cosine algorithm is employed to locate new food sources, thereby expanding the search scope and avoiding local optima. Lastly, a “madness factor” is added when identifying new sleeping burrows to further widen the search area and effectively circumvent local optima. Comparative analyses using benchmark functions demonstrate the improved algorithm’s superior convergence and stability. In this study, the LDMOA algorithm optimizes the weights and thresholds of the ELM to establish the LDMOA-ELM prediction model. Experimental forecasts utilizing data from China’s 2016 “The Electrician Mathematical Contest in Modeling” demonstrate that the LDMOA-ELM model significantly outperforms the original ELM model in terms of prediction error and accuracy.

Evaluation of the Potential of a Biocoagulant Produced from Prickly Pear Peel Waste Valorization for Wastewater Treatment

This study evaluated the potential of a biocoagulant produced from prickly pear peel waste valorization and its use as a biocoagulant aid mixed with aluminum sulfate to remove turbidity in domestic wastewater. A central composite design (CCD) and a simplex lattice design (SLD) of two components (biocoagulant and aluminum sulfate) were developed to determine the optimal doses and pH of the biocoagulant and optimal mixing proportions. Both designs optimized the coagulation process from an analysis of variance to fit the experimental data to mathematical models and an optimization analysis to obtain the highest percentage of turbidity removal. The results showed that a water pH of 4 and a biocoagulant dose of 100 mg/L are optimal conditions for a turbidity removal of 76.1%. The potential decreases to 51.7% when the wastewater pH is maintained at 7.8 and a dose of 250 mg/L is used. This efficiency could be increased to 58.2% by using a mixture with optimal proportions of 30% biocoagulant and 70% aluminum sulfate. The experimental data were fitted to two quadratic models, estimating model prediction errors of 0.42% and 2.34%, respectively. Therefore, these results support the valorization of prickly pear peel waste to produce a biocoagulant, which could be used in acid and alkaline wastewater or as a biocoagulant aid mixed with aluminum sulfate.

Atmospheric Chemistry Surrogate Modeling With Sparse Identification of Nonlinear Dynamics

AbstractModeling atmospheric chemistry is computationally expensive and limits the widespread use of chemical transport models. This computational cost arises from solving high‐dimensional systems of stiff differential equations. Previous work has demonstrated the promise of machine learning (ML) to accelerate air quality model simulations but has suffered from numerical instability during long‐term simulations. This may be because previous ML‐based efforts have relied on explicit Euler time integration—which is known to be unstable for stiff systems—and have used neural networks which are prone to overfitting. We hypothesize that parsimonious models combined with modern numerical integration techniques can overcome this limitation. Using a small‐scale mechanism to explore the potential of these methods, we have created a machine‐learned surrogate by (a) reducing dimensionality using singular value decomposition to create an interpretably‐compressed low‐dimensional latent space and (b) using Sparse Identification of Nonlinear Dynamics (SINDy) to create a differential‐equation‐based representation of the underlying dynamics in the compressed latent space with reduced stiffness. The root mean square error of ML model prediction for ozone concentration over 9 days is 37.8% of the root mean square concentration across all simulations in our testing data set. The surrogate model is 10× faster with 12× fewer integration timesteps compared to reference model and is numerically stable in all tested simulations. Overall, we find that SINDy can be used to create fast, stable, and accurate surrogates of a simple photochemical mechanism. In future work, we will explore the application of this method to more detailed mechanisms and their use in large‐scale simulations.

Subsetting reduces the error of MIR spectroscopy models for soil organic carbon prediction in the U.S. Great Plains

High demand for soil organic carbon data to support soil health and climate change mitigation efforts must be met with rapid, accurate, and inexpensive measurement methods. Mid-infrared spectroscopy shows promise as an alternative to conventional soil carbon analysis; however, its practicality depends on the construction and efficient use of soil spectral libraries. Subsetting, a calibration optimization technique, has potential to reduce model prediction errors. Nevertheless, the effectiveness of different subsetting criteria has yet to be well explored. This study assessed whether several subsetting criteria would result in calibration models with reduced error in the prediction of soil organic carbon content compared to models constructed from a full spectral dataset. A mid-infrared spectral library composed of Nebraska and Kansas soil samples was subset by (i) a nested wetland criterion, (ii) soilscapes, (iii) presence or absence of carbonates, and (iv) a combination of soilscape and carbonates. Partial least squares regression was used to construct all calibration models. Predictive performance of the subset models was compared to that of their corresponding full set model using several statistical metrics. Subsetting by wetlands reduced model error by 22 and 56 %. Subsetting by soilscape yielded a 13 to 55 % reduction in model error, while presence or absence of carbonates reduced model error by 21 and 46 %. Five of the eight combination soilscape and carbonate subset models reduced model error by 14 to 51 %. Overall, subsetting by soilscape or carbonate presence proved effective in improving model performance, with combination subsets proving beneficial under specific calibration set conditions.

Study on flow pattern transition of cuttings transport in extended-reach drilling

Inadequate hole cleaning is one of the major challenges in extended-reach drilling, especially prominent in large-sized hole sections, such as 121/4-in. section. As the drilling parameters including flow rate, ROP and rotational speed change, the cuttings flow pattern converts between fully suspended flow, two-layer flow, and three-layer flow. However, most existing cuttings transport models only study the problem of cuttings transport under a certain flow pattern and do not consider the dynamic conversion of different cuttings flow patterns. Pipe rotation plays a key role in cuttings transport in rotary drilling, but the existing studies have not fully revealed the effect of pipe rotation on the conversion of cuttings transport patterns. In this paper, a stirring diffusion factor is presented to quantify the effect of pipe rotation on cuttings transport, and the mathematical expression of the stirring diffusion factor is obtained through numerical simulations and the nonlinear regression method. Secondly, the critical conditions for the conversion of cuttings flow patterns are obtained by combining mechanical analysis and numerical simulation results, clarifying the applicable scope of different cuttings flow patterns. According to the comparison with experimental results, the prediction error of the new model considering the conversion of cuttings flow patterns is not higher than 11.1%, and the minimum prediction error is only 3.07%. Thirdly, the flow pattern maps for cuttings transport under different drilling parameters are drawn to provide guidance for the identification of cuttings transport pattern and optimization of drilling parameters. Finally, the flow pattern maps are applied to a case study of an extended-reach drilling. The results show that there is a delay effect in the conversion of cuttings flow patterns due to the nonlinear characteristics of annular geometry. Furthermore, the effects of flow rate, ROP and rotational speed on cuttings transport patterns are not independent, but mutually reinforcing and constraining each other. Inappropriate selection of drilling parameters will result in three-layer flow of cuttings transport, and then pipe sticking is likely to occur. The cuttings transport pattern can be converted from three-layer flow to two-layer flow by optimizing drilling parameters, which can significantly improve the hole cleaning effect.

Developing Parametric and Nonparametric Models for Model-Informed Precision Dosing: A Quality Improvement Effort in Vancomycin for Patients With Obesity.

Both parametric and nonparametric methods have been proposed to support model-informed precision dosing (MIPD). However, which approach leads to better models remains uncertain. Using open-source software, these 2 statistical approaches for model development were compared using the pharmacokinetics of vancomycin in a challenging subpopulation of class 3 obesity. Patients on vancomycin at the University of Vermont Medical Center from November 1, 2021, to February 14, 2023, were entered into the MIPD software. The inclusion criteria were body mass index (BMI) of at least 40 kg/m 2 and 1 or more vancomycin levels. A parametric model was created using nlmixr2/NONMEM, and a nonparametric model was created using Pmetrics. Then, a priori and a posteriori predictions were evaluated using the normalized root mean squared error (nRMSE) for precision and the mean percentage error (MPE) for bias. The parametric model was evaluated in a simulated MIPD context using an external validation dataset. In total, 83 patients were included in the model development, with a median age of 56.6 years (range: 24-89 years), and a median BMI of 46.3 kg/m 2 (range: 40-70.3 kg/m 2 ). Both parametric and nonparametric models were 2-compartmental, with creatinine clearance and fat-free mass as covariates to clearance and volume parameters, respectively. The a priori MPE and nRMSE for the parametric versus nonparametric models were -6.3% versus 2.69% and 27.2% versus 30.7%, respectively. The a posteriori MPE and RMSE were 0.16% and 0.84%, and 13.8% and 13.1%. The parametric model matched or outperformed previously published models on an external validation dataset (n = 576 patients). Minimal differences were found in the model structure and predictive error between the parametric and nonparametric approaches for modeling vancomycin class 3 obesity. However, the parametric model outperformed several other models, suggesting that institution-specific models may improve pharmacokinetics management.

Learning-Based Vehicle State Estimation Using Gaussian Process Regression Combined with Extended Kalman Filter

In order to address the challenge of accurate vehicle state estimation, especially in highly nonlinear and complex operating conditions, this paper proposes a learning-based method for vehicle state estimation. To achieve this, a novel theoretical framework is constructed that combines the model-based Extended Kalman Filter (EKF) with Gaussian process regression (GPR). By establishing a dynamic model of the vehicle and describing the errors using machine learning techniques, the sources of state estimation errors are analyzed. Furthermore, by considering the model uncertainty, an error prediction model is developed through GPR learning from real-world vehicle data. Combined with EKF, the proposed method enables high-precision online estimation of vehicle state. To validate the proposed method, tests are performed on a real vehicle test platform equipped with high-precision sensors and data acquisition equipment, under two different operating conditions: emergency and normal. The results are compared with those obtained using EKF and State Observer, which demonstrates a more accurate and adaptable vehicle state estimation, and provides a method for quantitatively describing the confidence interval of vehicle state estimation result.

Dynamic error prediction and link strain feedback control for a novel heavy load multi-DOF envelope forming machine

Thin walled and high rib components (TWHRC) are widely used for rocket cabins, aircraft wings, panels and frames in aerospace equipment as critical load bearing structures. This paper is aimed to develop a novel multi-DOF envelope forming machine (MEFM) with parallel kinematic mechanism (PKM) to realize efficient manufacturing of TWHRC with high performance. Since the heavy load MEFM with PKM has many total differences with the light load PKM, such as dynamic error and precision control, this paper conducts research on dynamic error prediction modelling and precision control method for this novel heavy load MEFM.Firstly, the configuration of MEFM is developed and the kinematic model of MEFM is established. Then, a RFC (rigid-flexible coupling) and RFE (rigid finite element) coupling dynamic model of MEFM is established considering time-variation of kinematic configuration caused by large deformation of all the kinematic elements. Based on the dynamic model, the dynamic deformation and error of MEFM are predicted and their generation mechanisms are revealed. It is found that the linear compression deformation of two-force links is the largest among all the kinematic elements, and it is the most sensitive to the heavy forming load and the most insensitive to the installation error of strain gauges, thus suiting for the feedback of large deformation of MEFM. Based on the above findings, a novel link strain feedback and dynamic error compensation control method is proposed. Finally, experiments under different control methods are conducted, showing that the motion precision of envelope tool and forming precision of formed components under the proposed control method are remarkably improved compared with that under the traditional control method (The geometric deviation of formed components under the proposed control method is remarkably reduced to −18 ∼ 32um from −64 ∼ 54um under the traditional control method). The experimental results verify that the proposed dynamic error prediction model and link strain feedback control method for heavy load MEFM are reasonable.

Bi-scale car-following model calibration based on corridor-level trajectory

The precise estimation of macroscopic traffic parameters, such as travel time and fuel consumption, is essential for the optimization of traffic management systems. Despite its importance, the comprehensive acquisition of vehicle trajectory data for the calculation of these macroscopic measures presents a challenge. To bridge this gap, this study aims to calibrate car-following models capable of predicting both microscopic measures and macroscopic measures. We conduct a numerical analysis to trace the cumulative process of model prediction errors across various measurements, and our findings indicate that macroscopic measures encapsulate the accumulation of model errors. By incorporating macroscopic measures into vehicle model calibration, we can mitigate the impact of noise on microscopic data measurements. We compare three car-following model calibration methods: MiC (using microscopic measurements), MaC (using macroscopic measurements), and BiC (using both microscopic and macroscopic measurements)—utilizing real-world trajectory data. The BiC method emerges as the most successful in reconstructing vehicle trajectories and accurately estimating travel time and fuel consumption, whereas the MiC method leads to overfitting and inaccurate macro-measurement predictions. This study underscores the importance of bi-scale calibration for precise traffic and energy consumption predictions, laying the groundwork for future research aimed at enhancing traffic management strategies.

Understanding and Mitigating Bias in Imaging Artificial Intelligence.

Artificial intelligence (AI) algorithms are prone to bias at multiple stages of model development, with potential for exacerbating health disparities. However, bias in imaging AI is a complex topic that encompasses multiple coexisting definitions. Bias may refer to unequal preference to a person or group owing to preexisting attitudes or beliefs, either intentional or unintentional. However, cognitive bias refers to systematic deviation from objective judgment due to reliance on heuristics, and statistical bias refers to differences between true and expected values, commonly manifesting as systematic error in model prediction (ie, a model with output unrepresentative of real-world conditions). Clinical decisions informed by biased models may lead to patient harm due to action on inaccurate AI results or exacerbate health inequities due to differing performance among patient populations. However, while inequitable bias can harm patients in this context, a mindful approach leveraging equitable bias can address underrepresentation of minority groups or rare diseases. Radiologists should also be aware of bias after AI deployment such as automation bias, or a tendency to agree with automated decisions despite contrary evidence. Understanding common sources of imaging AI bias and the consequences of using biased models can guide preventive measures to mitigate its impact. Accordingly, the authors focus on sources of bias at stages along the imaging machine learning life cycle, attempting to simplify potentially intimidating technical terminology for general radiologists using AI tools in practice or collaborating with data scientists and engineers for AI tool development. The authors review definitions of bias in AI, describe common sources of bias, and present recommendations to guide quality control measures to mitigate the impact of bias in imaging AI. Understanding the terms featured in this article will enable a proactive approach to identifying and mitigating bias in imaging AI. Published under a CC BY 4.0 license. Test Your Knowledge questions for this article are available in the supplemental material. See the invited commentary by Rouzrokh and Erickson in this issue.

Reviving Bistable Perception in Patients With Depression by Decreasing the Overestimation of Prior Precision.

Slower perceptual alternations, a notable perceptual effect observed in psychiatric disorders, can be alleviated by antidepressant therapies that affect serotonin levels in the brain. While these phenomena have been well documented, the underlying neurocognitive mechanisms remain to be elucidated. Our study bridges this gap by employing a computational cognitive approach within a Bayesian predictive coding framework to explore these mechanisms in depression. We fitted a prediction error (PE) model to behavioral data from a binocular rivalry task, uncovering that significantly higher initial prior precision and lower PE led to a slower switch rate in patients with depression. Furthermore, serotonin-targeting antidepressant treatments significantly decreased the prior precision and increased PE, both of which were predictive of improvements in the perceptual alternation rate of depression patients. These findings indicated that the substantially slower perception switch rate in patients with depression was caused by the greater reliance on top-down priors and that serotonin treatment's efficacy was in its recalibration of these priors and enhancement of PE. Our study not only elucidates the cognitive underpinnings of depression, but also suggests computational modeling as a potent tool for integrating cognitive science with clinical psychology, advancing our understanding and treatment of cognitive impairments in depression.

Calculation model of depth under quantification of surface morphology of dry shrinkage cracks in red clay

Cracks commonly form in clay owing to water loss, negatively affecting strength and hydraulic properties of soil, leading to rock and soil damage including landslides and dam instability. Understanding the changing patterns of the shrinkage crack depth is crucial for exploring slope infiltration patterns and determining key factors for slope stability. This study suggests a relationship between the surface and depth variations during cracking, and investigates the influence of the surface roughness on shrinkage cracks in saturated red clay. Regression analysis and machine learning algorithms were used to establish a depth prediction model. (1) Based on the surface morphology of shrinkage cracks, they are categorized into three types: linear, curved, and inflectional (linear and curved are the main types). (2) The surface roughness and development rate of surface cracks were positively correlated. (3) Based on the geometric parameters of the crack surface, a prediction model (error within 15%) for crack depth was established using multivariate nonlinear regression, providing a reference for the initial assessment of crack depth. These results provide a better understanding of the development patterns of shrinkage cracks, whose influence on slope stability is critical for mitigating the risks associated with slope instability and other soil failures.

Reconstruction of elasticity modulus distribution base on semi-supervised neural network

Accurate reconstruction of tissue elasticity modulus distribution has always been an important challenge in ultrasound elastography. Considering that existing deep learning-based supervised reconstruction methods only use simulated displacement data with random noise in training, which cannot fully provide the complexity and diversity brought by in-vivo ultrasound data, this study introduces the use of displacement data obtained by tracking in-vivo ultrasound radio frequency signals (i.e., real displacement data) during training, employing a semi-supervised approach to enhance the prediction accuracy of the model. Experimental results indicate that in phantom experiments, the semi-supervised model augmented with real displacement data provides more accurate predictions, with mean absolute errors and mean relative errors both around 3%, while the corresponding data for the fully supervised model are around 5%. When processing real displacement data, the area of prediction error of semi-supervised model was less than that of fully supervised model. The findings of this study confirm the effectiveness and practicality of the proposed approach, providing new insights for the application of deep learning methods in the reconstruction of elastic distribution from in-vivo ultrasound data.