The Effect of Training Sample Size on the Prediction of White Matter Hyperintensity Volume in a Healthy Population Using BIANCA

Introduction: White matter hyperintensities of presumed vascular origin (WMH) are an important magnetic resonance imaging marker of cerebral small vessel disease and are associated with cognitive decline, stroke, and mortality. Their relevance in healthy individuals, however, is less clear. This is partly due to the methodological challenge of accurately measuring rare and small WMH with automated segmentation programs. In this study, we tested whether WMH volumetry with FMRIB software library v6.0 (FSL; https://fsl.fmrib.ox.ac.uk/fsl/fslwiki) Brain Intensity AbNormality Classification Algorithm (BIANCA), a customizable and trainable algorithm that quantifies WMH volume based on individual data training sets, can be optimized for a normal aging population.Methods: We evaluated the effect of varying training sample sizes on the accuracy and the robustness of the predicted white matter hyperintensity volume in a population (n = 201) with a low prevalence of confluent WMH and a substantial proportion of participants without WMH. BIANCA was trained with seven different sample sizes between 10 and 40 with increments of 5. For each sample size, 100 random samples of T1w and FLAIR images were drawn and trained with manually delineated masks. For validation, we defined an internal and external validation set and compared the mean absolute error, resulting from the difference between manually delineated and predicted WMH volumes for each set. For spatial overlap, we calculated the Dice similarity index (SI) for the external validation cohort.Results: The study population had a median WMH volume of 0.34 ml (IQR of 1.6 ml) and included n = 28 (18%) participants without any WMH. The mean absolute error of the difference between BIANCA prediction and manually delineated masks was minimized and became more robust with an increasing number of training participants. The lowest mean absolute error of 0.05 ml (SD of 0.24 ml) was identified in the external validation set with a training sample size of 35. Compared to the volumetric overlap, the spatial overlap was poor with an average Dice similarity index of 0.14 (SD 0.16) in the external cohort, driven by subjects with very low lesion volumes.Discussion: We found that the performance of BIANCA, particularly the robustness of predictions, could be optimized for use in populations with a low WMH load by enlargement of the training sample size. Further work is needed to evaluate and potentially improve the prediction accuracy for low lesion volumes. These findings are important for current and future population-based studies with the majority of participants being normal aging people.

A Proposal for Parameter-Free Surrogate Building Algorithm Using Artificial Neural Networks

Surrogate models, capable of emulating the robust first principle based models, facilitate the online implementation of computationally expensive industrial process optimization. However, the heuristic estimation of parameters governing the surrogate building often renders them erroneous or under-trained. Current work aims at presenting a novel parameter free surrogate building approach, specifically focusing on Artificial Neural Networks. The proposed algorithm implements Sobol sampling plan and intelligently designs the configuration of network with simultaneous estimation of optimal transfer function and training sample size to prevent overfitting and enabling maximum prediction accuracy. A novel Sample Size Determination algorithm based on a potential concept of hypercube sampling technique adds to the speed of surrogate building algorithm, thereby assuring faster convergence. Surrogates models for a highly nonlinear industrial sintering process constructed using the novel algorithm resulted in 7 times faster optimization.

Enhancing Handover for 5G Mobile Networks using Jump Markov Linear System and Deep Reinforcement Learning

The fifth Generation (5G) mobile networks use millimeter Waves (mmWaves) to offer giga bit data rates. However, unlike microwaves, mmWave links are prone to user and topographic dynamics. They easily get blocked and end up forming irregular cell patterns for 5G. This in turn cause too early, too late, or wrong handoffs (HOs). To mitigate HO challenges, sustain connectivity and avert unnecessary HO, we propose a HO scheme based on Jump Markov Linear System (JMLS) and Deep Reinforcement Learning (DRL). JMLS is widely known to account for abrupt changes in system dynamics. DRL likewise emerges as an artificial intelligence technique for learning highly dimensional and time-varying behaviors. We combine the two techniques to account for time-varying, abrupt, and irregular changes in mmWave link behaviour by predicting likely deterioration patterns of target links. The prediction is optimized by meta training techniques that also reduces training sample size. Thus, the JMLS-DRL platform formulates intelligent and versatile HO policies for 5G. Results show our proposed prediction scheme about target link behavior post HO to be highly reliable. The scheme also averts unnecessary HOs thus ably supports longer dew time.

Emulation of 2D Hydrodynamic Flood Simulations at Catchment Scale Using ANN and SVR

Two-dimensional (2D) hydrodynamic models are one of the most widely used tools for flood modeling practices and risk estimation. The 2D models provide accurate results; however, they are computationally costly and therefore unsuitable for many real time applications and uncertainty analysis that requires a large number of model realizations. Therefore, the present study aims to (i) develop emulators based on SVR and ANN as an alternative for predicting the 100-year flood water level, (ii) improve the performance of the emulators through dimensionality reduction techniques, and (iii) assess the required training sample size to develop an accurate emulator. Our results indicate that SVR based emulator is a fast and reliable alternative that can predict the water level accurately. Moreover, the performance of the models can improve by identifying the most influencing input variables and eliminating redundant inputs from the training process. The findings in this study suggest that the training data size equal to 70% (or more) of data results in reliable and accurate predictions.

ASSESSMENT OF THE ELECTRICAL EQUIPMENT INSULATION STATE USING THE GRADIENT BOOSTING ALGORITHM

The creation of analytical software products aimed at assessing the electrical equipment state has become a priority in the development of diagnostics in the power industry. The artificial intelligence methods are useful for this problem-solving. In the article, we propose a method for analyzing the monitoring data of partial discharges in the insulation of electrical equipment using machine-learning technologies. An analytical assessment of the partial discharges characteristics allows us to conclude on the insulation state of the object. It is proposed to use integrated diagnostic parameters, such as partial discharges intensity – the maximum measured value of the apparent charge of a single, repetitive and regular partial discharges. The total sample is characterized by an imbalance, which is typical for technical diagnostics in general. Among machine learning algorithms, bagging and boosting have proven to be the most effective. The mathematical apparatus of gradient boosting is considered in the example of the most common algorithms GBM (Gradient Boosting Machine) and CatBoost. The model was created in the Python programming language. The model created on the basis of the CatBoost algorithm was used for assessing the condition of the oil insulation of power transformers. The model’s accuracy of 68.85% was achieved after optimizing the parameters of the CatBoost algorithm. The article concluded that it is necessary to increase the training sample size and improve its balance. It is inadvisable to interpret the predicted data in the field of diagnostics parameters at the available accuracy of the model’s wok.

Improving the Generalizability of Infantile Cataracts Detection via Deep Learning-Based Lens Partition Strategy and Multicenter Datasets

Infantile cataract is the main cause of infant blindness worldwide. Although previous studies developed artificial intelligence (AI) diagnostic systems for detecting infantile cataracts in a single center, its generalizability is not ideal because of the complicated noises and heterogeneity of multicenter slit-lamp images, which impedes the application of these AI systems in real-world clinics. In this study, we developed two lens partition strategies (LPSs) based on deep learning Faster R-CNN and Hough transform for improving the generalizability of infantile cataracts detection. A total of 1,643 multicenter slit-lamp images collected from five ophthalmic clinics were used to evaluate the performance of LPSs. The generalizability of Faster R-CNN for screening and grading was explored by sequentially adding multicenter images to the training dataset. For the normal and abnormal lenses partition, the Faster R-CNN achieved the average intersection over union of 0.9419 and 0.9107, respectively, and their average precisions are both > 95%. Compared with the Hough transform, the accuracy, specificity, and sensitivity of Faster R-CNN for opacity area grading were improved by 5.31, 8.09, and 3.29%, respectively. Similar improvements were presented on the other grading of opacity density and location. The minimal training sample size required by Faster R-CNN is determined on multicenter slit-lamp images. Furthermore, the Faster R-CNN achieved real-time lens partition with only 0.25 s for a single image, whereas the Hough transform needs 34.46 s. Finally, using Grad-Cam and t-SNE techniques, the most relevant lesion regions were highlighted in heatmaps, and the high-level features were discriminated. This study provides an effective LPS for improving the generalizability of infantile cataracts detection. This system has the potential to be applied to multicenter slit-lamp images.