Early prediction of mortality at sepsis diagnosis time in critically ill patients by using interpretable machine learning.

This study applied machine learning for the early prediction of 30-day mortality at sepsis diagnosis time in critically ill patients. Retrospective study using data collected from the Medical Information Mart for Intensive Care IV database. The data of the patient cohort was divided on the basis of the year of hospitalization, into training (2008-2013), validation (2014-2016), and testing (2017-2019) datasets. 24,377 patients with the sepsis diagnosis time < 24 h after intensive care unit (ICU) admission were included. A gradient boosting tree-based algorithm (XGBoost) was used for training the machine learning model to predict 30-day mortality at sepsis diagnosis time in critically ill patients. Model performance was measured in both discrimination and calibration aspects. The model was interpreted using the SHapley Additive exPlanations (SHAP) module. The 30-day mortality rate of the testing dataset was 17.9%, and 39 features were selected for the machine learning model. Model performance on the testing dataset achieved an area under the receiver operating characteristic curve (AUROC) of 0.853 (95% CI 0.837-0.868) and an area under the precision-recall curves of 0.581 (95% CI 0.541-0.619). The calibration plot for the model revealed a slope of 1.03 (95% CI 0.94-1.12) and intercept of 0.14 (95% CI 0.04-0.25). The SHAP revealed the top three most significant features, namely age, increased red blood cell distribution width, and respiratory rate. Our study demonstrated the feasibility of using the interpretable machine learning model to predict mortality at sepsis diagnosis time.

ABC-Norm Regularization for Fine-Grained and Long-Tailed Image Classification.

Image classification for real-world applications often involves complicated data distributions such as fine-grained and long-tailed. To address the two challenging issues simultaneously, we propose a new regularization technique that yields an adversarial loss to strengthen the model learning. Specifically, for each training batch, we construct an adaptive batch prediction (ABP) matrix and establish its corresponding adaptive batch confusion norm (ABC-Norm). The ABP matrix is a composition of two parts, including an adaptive component to class-wise encode the imbalanced data distribution, and the other component to batch-wise assess the softmax predictions. The ABC-Norm leads to a norm-based regularization loss, which can be theoretically shown to be an upper bound for an objective function closely related to rank minimization. By coupling with the conventional cross-entropy loss, the ABC-Norm regularization could introduce adaptive classification confusion and thus trigger adversarial learning to improve the effectiveness of model learning. Different from most of state-of-the-art techniques in solving either fine-grained or long-tailed problems, our method is characterized with its simple and efficient design, and most distinctively, provides a unified solution. In the experiments, we compare ABC-Norm with relevant techniques and demonstrate its efficacy on several benchmark datasets, including (CUB-LT, iNaturalist2018); (CUB, CAR, AIR); and (ImageNet-LT), which respectively correspond to the real-world, fine-grained, and long-tailed scenarios.

Toward automatic beam angle selection for pencil-beam scanning proton liver treatments: A deep learning-based approach.

BACKGROUND: Dose deposition characteristics of proton radiation can be advantageous over photons. Proton treatment planning, however, poses additional challenges for the planners. Proton therapy is usually delivered with only a small number of beam angles, and the quality of a proton treatment plan is largely determined by the beam angles employed. Finding the optimal beam angles for a proton treatment plan requires time and experience, motivating the investigation of automatic beam angle selection methods. PURPOSE: A deep learning-based approach to automatic beam angle selection is proposed for the proton pencil-beam scanning treatment planning of liver lesions. METHODS: We cast beam-angle selection as a multi-label classification problem. To account for angular boundary discontinuity, the underlying convolution neural network is trained with the proposed Circular Earth Mover's Distance-based regularization and multi-label circular-smooth label technique. Furthermore, an analytical algorithm emulating proton treatment planners' clinical practice is employed in post-processing to improve the output of the model. Forty-nine patients that received proton liver treatments between 2017 and 2020 were randomly divided into training (n = 31), validation (n = 7), and test sets (n = 11). AI-selected beam angles were compared with those angles selected by human planners, and the dosimetric outcome was investigated by creating plans using knowledge-based treatment planning. RESULTS: For 7 of the 11 cases in the test set, AI-selected beam angles agreed with those chosen by human planners to within 20° (median angle difference = 10°; mean = 18.6°). Moreover, out of the total 22 beam angles predicted by the model, 15 (68%) were within 10° of the human-selected angles. The high correlation in beam angles resulted in comparable dosimetric statistics between proton treatment plans generated using AI- and human-selected angles. For the cases with beam angle differences exceeding 20°, the dosimetric analysis showed similar plan quality although with different emphases on organ-at-risk sparing. CONCLUSIONS: This pilot study demonstrated the feasibility of a novel deep learning-based beam angle selection technique. Testing on liver cancer patients showed that the resulting plans were clinically viable with comparable dosimetric quality to those using human-selected beam angles. In tandem with auto-contouring and knowledge-based treatment planning tools, the proposed model could represent a pathway for nearly fully automated treatment planning in proton therapy.

Determining Carina and Clavicular Distance-Dependent Positioning of Endotracheal Tube in Critically Ill Patients: An Artificial Intelligence-Based Approach.

Early and accurate prediction of endotracheal tube (ETT) location is pivotal for critically ill patients. Automatic and timely detection of faulty ETT locations from chest X-ray images may avert patients' morbidity and mortality. Therefore, we designed convolutional neural network (CNN)-based algorithms to evaluate ETT position appropriateness relative to four detected key points, including tracheal tube end, carina, and left/right clavicular heads on chest radiographs. We estimated distances from the tube end to tracheal carina and the midpoint of clavicular heads. A DenseNet121 encoder transformed images into embedding features, and a CNN-based decoder generated the probability distributions. Based on four sets of tube-to-carina distance-dependent parameters (i.e., (i) 30-70 mm, (ii) 30-60 mm, (iii) 20-60 mm, and (iv) 20-55 mm), corresponding models were generated, and their accuracy was evaluated through the predicted L1 distance to ground-truth coordinates. Based on tube-to-carina and tube-to-clavicle distances, the highest sensitivity, and specificity of 92.85% and 84.62% respectively, were revealed for 20-55 mm. This implies that tube-to-carina distance between 20 and 55 mm is optimal for an AI-based key point appropriateness detection system and is empirically comparable to physicians' consensus.

ezGeno: an automatic model selection package for genomic data analysis.

MOTIVATION: To facilitate the process of tailor-making a deep neural network for exploring the dynamics of genomic DNA, we have developed a hands-on package called ezGeno. ezGeno automates the search process of various parameters and network structures and can be applied to any kind of 1D genomic data. Combinations of multiple abovementioned 1D features are also applicable. RESULTS: For the task of predicting TF binding using genomic sequences as the input, ezGeno can consistently return the best performing set of parameters and network structure, as well as highlight the important segments within the original sequences. For the task of predicting tissue-specific enhancer activity using both sequence and DNase feature data as the input, ezGeno also regularly outperforms the hand-designed models. Furthermore, we demonstrate that ezGeno is superior in efficiency and accuracy compared to the one-layer DeepBind model and AutoKeras, an open-source AutoML package. AVAILABILITY AND IMPLEMENTATION: The ezGeno package can be freely accessed at https://github.com/ailabstw/ezGeno. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Multiple kernel learning for dimensionality reduction.

In solving complex visual learning tasks, adopting multiple descriptors to more precisely characterize the data has been a feasible way for improving performance. The resulting data representations are typically high-dimensional and assume diverse forms. Hence, finding a way of transforming them into a unified space of lower dimension generally facilitates the underlying tasks such as object recognition or clustering. To this end, the proposed approach (termed MKL-DR) generalizes the framework of multiple kernel learning for dimensionality reduction, and distinguishes itself with the following three main contributions: first, our method provides the convenience of using diverse image descriptors to describe useful characteristics of various aspects about the underlying data. Second, it extends a broad set of existing dimensionality reduction techniques to consider multiple kernel learning, and consequently improves their effectiveness. Third, by focusing on the techniques pertaining to dimensionality reduction, the formulation introduces a new class of applications with the multiple kernel learning framework to address not only the supervised learning problems but also the unsupervised and semi-supervised ones.

Regularized background adaptation: a novel learning rate control scheme for gaussian mixture modeling.

To model a scene for background subtraction, Gaussian mixture modeling (GMM) is a popular choice for its capability of adaptation to background variations. However, GMM often suffers from a tradeoff between robustness to background changes and sensitivity to foreground abnormalities and is inefficient in managing the tradeoff for various surveillance scenarios. By reviewing the formulations of GMM, we identify that such a tradeoff can be easily controlled by adaptive adjustments of the GMM's learning rates for image pixels at different locations and of distinct properties. A new rate control scheme based on high-level feedback is then developed to provide better regularization of background adaptation for GMM and to help resolving the tradeoff. Additionally, to handle lighting variations that change too fast to be caught by GMM, a heuristic rooting in frame difference is proposed to assist the proposed rate control scheme for reducing false foreground alarms. Experiments show the proposed learning rate control scheme, together with the heuristic for adaptation of over-quick lighting change, gives better performance than conventional GMM approaches.

Real-time tracking using trust-region methods.

Optimization methods based on iterative schemes can be divided into two classes: line-search methods and trust-region methods. While line-search techniques are commonly found in various vision applications, not much attention is paid to trust-region ones. Motivated by the fact that line-search methods can be considered as special cases of trust-region methods, we propose to establish a trust-region framework for real-time tracking. Our approach is characterized by three key contributions. First, since a trust-region tracking system is more effective, it often yields better performances than the outcomes of other trackers that rely on iterative optimization to perform tracking, e.g., a line-search-based mean-shift tracker. Second, we have formulated a representation model that uses two coupled weighting schemes derived from the covariance ellipse to integrate an object's color probability distribution and edge density information. As a result, the system can address rotation and nonuniform scaling in a continuous space, rather than working on some presumably possible discrete values of rotation angle and scale. Third, the framework is very flexible in that a variety of distance functions can be adapted easily. Experimental results and comparative studies are provided to demonstrate the efficiency of the proposed method.