ORIENTATE: automated machine learning classifiers for oral health prediction and research.

BACKGROUND: The application of data-driven methods is expected to play an increasingly important role in healthcare. However, a lack of personnel with the necessary skills to develop these models and interpret its output is preventing a wider adoption of these methods. To address this gap, we introduce and describe ORIENTATE, a software for automated application of machine learning classification algorithms by clinical practitioners lacking specific technical skills. ORIENTATE allows the selection of features and the target variable, then automatically generates a number of classification models and cross-validates them, finding the best model and evaluating it. It also implements a custom feature selection algorithm for systematic searches of the best combination of predictors for a given target variable. Finally, it outputs a comprehensive report with graphs that facilitates the explanation of the classification model results, using global interpretation methods, and an interface for the prediction of new input samples. Feature relevance and interaction plots provided by ORIENTATE allow to use it for statistical inference, which can replace and/or complement classical statistical studies. RESULTS: Its application to a dataset with healthy and special health care needs (SHCN) children, treated under deep sedation, was discussed as case study. On the example dataset, despite its small size, the feature selection algorithm found a set of features able to predict the need for a second sedation with a f1 score of 0.83 and a ROC (AUC) of 0.92. Eight predictive factors for both populations were found and ordered by the relevance assigned to them by the model. A discussion of how to derive inferences from the relevance and interaction plots and a comparison with a classical study is also provided. CONCLUSIONS: ORIENTATE automatically finds suitable features and generates accurate classifiers which can be used in preventive tasks. In addition, researchers without specific skills on data methods can use it for the application of machine learning classification and as a complement to classical studies for inferential analysis of features. In the case study, a high prediction accuracy for a second sedation in SHCN children was achieved. The analysis of the relevance of the features showed that the number of teeth with pulpar treatments at the first sedation is a predictive factor for a second sedation.

Towards Outdoor Electromagnetic Field Exposure Mapping Generation Using Conditional GANs.

With the ongoing fifth-generation cellular network (5G) deployment, electromagnetic field exposure has become a critical concern. However, measurements are scarce, and accurate electromagnetic field reconstruction in a geographic region remains challenging. This work proposes a conditional generative adversarial network to address this issue. The main objective is to reconstruct the electromagnetic field exposure map accurately according to the environment's topology from a few sensors located in an outdoor urban environment. The model is trained to learn and estimate the propagation characteristics of the electromagnetic field according to the topology of a given environment. In addition, the conditional generative adversarial network-based electromagnetic field mapping is compared with simple kriging. Results show that the proposed method produces accurate estimates and is a promising solution for exposure map reconstruction.

Evaluation of blood glucose level control in type 1 diabetic patients using deep reinforcement learning.

Diabetes mellitus is a disease associated with abnormally high levels of blood glucose due to a lack of insulin. Combining an insulin pump and continuous glucose monitor with a control algorithm to deliver insulin is an alternative to patient self-management of insulin doses to control blood glucose levels in diabetes mellitus patients. In this work, we propose a closed-loop control for blood glucose levels based on deep reinforcement learning. We describe the initial evaluation of several alternatives conducted on a realistic simulator of the glucoregulatory system and propose a particular implementation strategy based on reducing the frequency of the observations and rewards passed to the agent, and using a simple reward function. We train agents with that strategy for three groups of patient classes, evaluate and compare it with alternative control baselines. Our results show that our method is able to outperform baselines as well as similar recent proposals, by achieving longer periods of safe glycemic state and low risk.

Opal: An open source ray-tracing propagation simulator for electromagnetic characterization.

Accurate characterization and simulation of electromagnetic propagation can be obtained by ray-tracing methods, which are based on a high frequency approximation to the Maxwell equations and describe the propagating field as a set of propagating rays, reflecting, diffracting and scattering over environment elements. However, this approach has been usually too computationally costly to be used in large and dynamic scenarios, but this situation is changing thanks the increasing availability of efficient ray-tracing libraries for graphical processing units. In this paper we present Opal, an electromagnetic propagation simulation tool implemented with ray-tracing on graphical processing units, which is part of the Veneris framework. Opal can be used as a stand-alone ray-tracing simulator, but its main strength lies in its integration with the game engine, which allows to generate customized 3D environments quickly and intuitively. We describe its most relevant features and provide implementation details, highlighting the different simulation types it supports and its extension possibilites. We provide application examples and validate the simulation on demanding scenarios, such as tunnels, where we compare the results with theoretical solutions and further discuss the tradeoffs between the simulation types and its performance.

Garment counting in a textile warehouse by means of a laser imaging system.

Textile logistic warehouses are highly automated mechanized places where control points are needed to count and validate the number of garments in each batch. This paper proposes and describes a low cost and small size automated system designed to count the number of garments by processing an image of the corresponding hanger hooks generated using an array of phototransistors sensors and a linear laser beam. The generated image is processed using computer vision techniques to infer the number of garment units. The system has been tested on two logistic warehouses with a mean error in the estimated number of hangers of 0.13%.

On the optimal identification of tag sets in time-constrained RFID configurations.

In Radio Frequency Identification facilities the identification delay of a set of tags is mainly caused by the random access nature of the reading protocol, yielding a random identification time of the set of tags. In this paper, the cumulative distribution function of the identification time is evaluated using a discrete time Markov chain for single-set time-constrained passive RFID systems, namely those ones where a single group of tags is assumed to be in the reading area and only for a bounded time (sojourn time) before leaving. In these scenarios some tags in a set may leave the reader coverage area unidentified. The probability of this event is obtained from the cumulative distribution function of the identification time as a function of the sojourn time. This result provides a suitable criterion to minimize the probability of losing tags. Besides, an identification strategy based on splitting the set of tags in smaller subsets is also considered. Results demonstrate that there are optimal splitting configurations that reduce the overall identification time while keeping the same probability of losing tags.