Automatic Strategy Research Articles

Classification of Chronic Venous Disorders using an Ensemble Optimization of Convolutional Neural Networks.

Chronic Venous Disorders (CVD) of lower limbs are one of the most prevalent medical conditions, affecting 35% of adults in Europe and North America. The early diagnosis of CVD is critical, however, the diagnosis relies on a visual recognition of the various venous disorders which is time- consuming and dependent on the physician's expertise. Thus, automatic strategies for the classification of the CVD severity are claimed. This paper proposed an automatic ensemble-based strategy of Deep Convolutional Neural Networks (DCNN) for the classification of CVDs severity from medical images. First, a clinical dataset containing 1376 images of patients' legs with CVD of 5 different levels of severity was constructed. Then, the constructed dataset was randomly split into training, testing, and validation datasets. Subsequently, a set of DCNN were individually applied to the images for classification. Finally, instead of a traditional voting ensemble strategy, extracted feature vectors from each DCNN were concatenated and fed into a new ensemble optimization network. Experiments showed that the proposed strategy achieved a classification with 93.8%, 93.4%, 92.4% of accuracy, precision, and recall, respectively. Moreover, compared to the traditional ensemble strategy, improvement in the accuracy of ~2% was registered. The proposed strategy showed to be accurate and robust for the diagnosis of CVD severity from medical images. Nevertheless, further research using an extensive clinical database is required to validate the potential of this strategy. Clinical Relevance- An automatic classification of CVD to reduce the probability of underdiagnoses and promote the treatment of CVD in the early stages.

Collision-free trajectory planning for multi-robot simultaneous motion in preforms weaving

AbstractIn this paper, an automatic obstacle avoidance trajectory planning strategy is proposed for the simultaneous motion of multi-robots, which perform anthropomorphic skill operations in a large curved preformed three-dimensional (3D) weaving environment with multiple obstacles and limited space, to eliminate tedious manual calibration work of robot path in engineering. Firstly, an Adaptive Goal-guided Rapidly-exploring Random Trees (AGG-RRT) algorithm is proposed, combined with the robot obstacle avoidance strategy, to search the discrete position of the collision-free path of the end-effector gradually from the starting point to the ending point. Then the optimization of the path is completed by bidirectional pruning of redundant nodes and cubic non-uniform rational B-spline (NURBS) curve fitting. And finally, the robot trajectory is interpolated based on S-shaped acceleration/deceleration planning to ensure smooth robot joint motion. The simulation results demonstrate the superiority of the AGG-RRT algorithm over the basic RRT algorithm and related improved algorithms in terms of search time and success rate. The simulated experiments also achieve the smooth trajectory planning of multiple robotic arms with the synchronous obstacle avoidance motion, which shows that the AGG-RRT algorithm is successfully applied and the collision-free trajectory planning strategy is effective.

A Regional Smoothing Block Sparse Bayesian Learning Method With Temporal Correlation for Channel Selection in P300 Speller.

The P300-based brain–computer interfaces (BCIs) enable participants to communicate by decoding the electroencephalography (EEG) signal. Different regions of the brain correspond to various mental activities. Therefore, removing weak task-relevant and noisy channels through channel selection is necessary when decoding a specific type of activity from EEG. It can improve the recognition accuracy and reduce the training time of the subsequent models. This study proposes a novel block sparse Bayesian-based channel selection method for the P300 speller. In this method, we introduce block sparse Bayesian learning (BSBL) into the channel selection of P300 BCI for the first time and propose a regional smoothing BSBL (RSBSBL) by combining the spatial distribution properties of EEG. The RSBSBL can determine the number of channels adaptively. To ensure practicality, we design an automatic selection iteration strategy model to reduce the time cost caused by the inverse operation of the large-size matrix. We verified the proposed method on two public P300 datasets and on our collected datasets. The experimental results show that the proposed method can remove the inferior channels and work with the classifier to obtain high-classification accuracy. Hence, RSBSBL has tremendous potential for channel selection in P300 tasks.

MS-FINDER Assisted in Understanding the Profile of Flavonoids in Temporal Dimension during the Fermentation of Pu-erh Tea.

Flavonoid represents a significant class of secondary metabolites in Pu-erh tea with benefits to human health. For a rapid and complete discovery of such compounds, we established a data mining workflow that integrates software MS-DIAL, MS-FINDER, and molecular networking analysis. As a result, 181 flavonoids were tentatively annotated including 22 first found in Pu-erh tea, and two of them were potentially new molecules. The dynamic alteration of these flavonoids during Pu-erh fermentation was further investigated. They all showed a trend of first increasing and then decreasing. Moreover, statistical analysis showed that the first to third pile turnings of the fermentation process had a greater impact on the changes of flavonoids. Partial metabolic pathways were proposed. This study provides a quick and automatic strategy for flavonoid profiling. The temporal dimension of flavonoids during fermentation may serve as a theoretical basis for Pu-erh tea manufacturing technology and study on substance foundation.

Employing automatic content recognition for teaching methodology analysis in classroom videos.

A teacher plays a pivotal role in grooming a society and paves way for its social and economic developments. Teaching is a dynamic role and demands continuous adaptation. A teacher adopts teaching techniques suitable for a certain discipline and a situation. A thorough, detailed, and impartial observation of a teacher is a desideratum for adaptation of an effective teaching methodology and it is a laborious exercise. An automatic strategy for analyzing a teacher's teaching methodology in a classroom environment is suggested in this work. The proposed strategy recognizes a teacher's actions in videos while he is delivering lectures. In this study, 3D CNN and Conv2DLSTM with time-distributed layers are used for experimentation. A range of actions are recognized for a complete classroom session during experimentation and the reported results are considered effective for analysis of a teacher's teaching technique.

The Cognitive–Creative Profiles of Insightful Problem Solvers: A Person‐Centered Insight Study

AbstractThis study aimed to portray insightful problem solvers. Participants solved classic insight problems and rebuses and completed cognitive and creativity tasks. Grouping the participants based on their success in solving insight problems of different modalities revealed that 95% of the spatial problem solvers succeeded in solving verbal ones (spatial verbal solvers) but not vice versa. Verbal problem solvers were not necessarily able to solve spatial ones. A latent class analysis (LCA) based on relationships among the cognitive and creativity variables confirmed these two types of insightful solvers. It portrayed two compatible classes of experts and successful solvers alongside two classes of low‐level solvers. The former classes excelled in solving rebuses and verbal problems, yet the experts outperformed the others in spatial problems. Both classes scored higher in spatial divergent thinking measures than the low‐level solvers. However, the experts benefited additionally from verbal divergent thinking and associative combination ability. Seemingly, should retrieval of diverse ideas be insufficient to solve the problem, the experts combine ideas by a novel association, thereby generating an insight. Moreover, individual differences in working memory indicated that experts favor associative–automatic strategy, whereas successful solvers utilize analytic strategy. These profiles illuminate “insight” in terms of cognitive and creative styles.

Automatic Load Model Selection Based on Machine Learning Algorithms

Technology development and decentralized operations create changes in conventional electric systems, where load modeling has been a challenge in dynamic analysis. Consequently, accurate dynamic load models are required to ensure the quality of the studies in current systems. This paper presents an automatic strategy based on clustering, classification, and optimization algorithms, to obtain the load models in the case of several system operating conditions. The obtained load models can be helpful for the planning and operation of electric power systems. The proposed approach validation is performed using the IEEE 14-bus test system, where high performance is obtained. The average obtained cross-validation error for the load models assigned to the 13 clusters of disturbances is 5.36 × 10^-3. The cross-validation error is used as a tolerance value to determine when an online assigned load model is suitable to represent the measured disturbance. The proposed tests show the strategy’s capabilities of defining the load model online, making this approach suitable for field applications.

Utilizing unsupervised learning, multi-view imaging, and CNN-based attention facilitates cost-effective wetland mapping

The combination of Unmanned/Unoccupied Aerial Vehicle (UAV) data and deep learning, especially convolutional neural networks (CNNs), offers robust new tools for precision land cover mapping. However, its successful application is highly dependent on local experiences that are rarely documented, resulting in practical limitations during implementation. Cost-effective deep learning frameworks for fast deployment are required. This study presents a deep learning adaptation framework, named Auto-UNet++, trying to streamline wetland mapping tasks (including training data labeling and organizing). The framework treats mapping tasks as an intact semantic segmentation pipeline and then integrates automatic strategies into each step to reduce human intervention. These automatic strategies are achieved by standard computer vision techniques, including multi-view (MV) imaging—highly overlapped UAV images over an area (for labeling/voting), unsupervised clustering (for labeling), multi-scale CNN (for feature extraction), and attention mechanism—a CNN design used to select informative features from input (for feature exploration/selection). The framework was tested on playa wetland mapping in the Rainwater Basin, Nebraska, USA, with multispectral UAV datasets. Generally, the multi-scale CNN mapping task achieved a high of 87% overall accuracy and over 90% accuracy in water delineation. The results indicate that the multi-view and attention strategies have the potential to improve segmentation performance, and together with unsupervised learning, save considerable labor/expertise. Interestingly, evidence shows that the band/scale attention (weight) is adaptively associated with the land cover percentages per input image, indicating spatial contexts captured. This finding highlights the potential usages of the attention rule in automatic feature exploration, selection, and model interpretation. The framework illustrating a highly automated deep learning deployment on small MV datasets facilitates cost-effective wetland cover mapping. Although limitations exist, the study demonstrated the possibility of where/how conventional segmentation pipelines can be improved in typical UAV wetland mapping tasks. The framework and findings are useful for similar applications (including non-UAV studies) that only have limited time, labor, and expertise to implement sophisticated semantic segmentation models.

Data-Driven Consumption Load Monitoring and Adjustment Strategy in Smart Grid

The enhancement of the intelligent construction of the power grid and widespread popularity of smart meters enable large amounts of electrical energy consumption data to be collected and analyzed. Based on the data, the energy provider gives a guiding price in the future periods to users. It encourages users to be more economical and smarter in the process of using electricity. By applying the social welfare model to equate demand and supply in every time interval, we gain the optimal prices and generation capacity. Nevertheless, the truth is that there is a great gap between the consumers’ booked electrical energy consumption and the optimal generation capacity, causing the power system overload and even outage. This article puts forward a novel automatic process control strategy in order to monitor the gap between the consumers’ booked electrical energy consumption and optimal generation capacity by using statistical method to predict the future one. When the predicted value exceeds the boundary, the energy provider adopts the changeable electricity price to stimulate consumers to adjust their electrical energy demands so that it can have smoothly actual electrical energy consumption. Our adjustment method is data-driven exponential function-based adjustment. Case study results show that the strategy can obtain small adjustment times, stable actual consumption load, and controllable prediction errors. Different from the linear monitoring and adjustment strategy, our approach obtains almost the same adjustment frequency, less standard deviation of residuals, and higher total social welfare and energy provider profit.

Neural-Fuzzy-Based Adaptive Sliding Mode Automatic Steering Control of Vision-based Unmanned Electric Vehicles

This paper presents a novel neural-fuzzy-based adaptive sliding mode automatic steering control strategy to improve the driving performance of vision-based unmanned electric vehicles with time-varying and uncertain parameters. Primarily, the kinematic and dynamic models which accurately express the steering behaviors of vehicles are constructed, and in which the relationship between the look-ahead time and vehicle velocity is revealed. Then, in order to overcome the external disturbances, parametric uncertainties and time-varying features of vehicles, a neural-fuzzy-based adaptive sliding mode automatic steering controller is proposed to supervise the lateral dynamic behavior of unmanned electric vehicles, which includes an equivalent control law and an adaptive variable structure control law. In this novel automatic steering control system of vehicles, a neural network system is utilized for approximating the switching control gain of variable structure control law, and a fuzzy inference system is presented to adjust the thickness of boundary layer in real-time. The stability of closed-loop neural-fuzzy-based adaptive sliding mode automatic steering control system is proven using the Lyapunov theory. Finally, the results illustrate that the presented control scheme has the excellent properties in term of error convergence and robustness.

A-Learn EvId: A Method for Identifying Evidence of Computer Programming Skills Through Automatic Source Code Assessment

Contextualized in the teaching of computer programming in Computing courses, this research investigates aspects and strategies for automatic source code assessment. Continuous on-time assessment of source codes produced by students is a challenging task for teachers. The literature presents different methods for automatic assessment of source code, mostly focusing on technical aspects, such as functional correctness assessment and error detection. This paper presents the A-Learn EvId method, having as the main characteristic its focus on the assessment of high-level skills instead of technical aspects. Automatically assessing high-level skills gives insights into the thought process students used to elaborate their responses, contributing to quality and timely feedback generation. The method is characterized by three fundamental steps: (1) inserting students' source code as input data; (2) identifying evidence of skills through automatic strategies; and (3) representing identified skills through a student model. The following contributions are highlighted: updating the state of the art on the topic; a set of 37 skills identifiable through 9 automatic source code assessment strategies; construction of datasets totaling 8651 source codes.

Early bradycardia detection and therapeutic interventions in preterm infant monitoring

In very preterm infants, cardio-respiratory events and associated hypoxemia occurring during early postnatal life have been associated with risks of retinopathy, growth alteration and neurodevelopment impairment. These events are commonly detected by continuous cardio-respiratory monitoring in neonatal intensive care units (NICU), through the associated bradycardia. NICU nurse interventions are mainly triggered by these alarms. In this work, we acquired data from 52 preterm infants during NICU monitoring, in order to propose an early bradycardia detector which is based on a decentralized fusion of three detectors. The main objective is to improve automatic detection under real-life conditions without altering performance with respect to that of a monitor commonly used in NICU. We used heart rate lower than 80 bpm during at least 10 sec to define bradycardia. With this definition we observed a high rate of false alarms (64%) in real-life and that 29% of the relevant alarms were not followed by manual interventions. Concerning the proposed detection method, when compared to current monitors, it provided a significant decrease of the detection delay of 2.9 seconds, without alteration of the sensitivity (97.6% vs 95.2%) and false alarm rate (63.7% vs 64.1%). We expect that such an early detection will improve the response of the newborn to the intervention and allow for the development of new automatic therapeutic strategies which could complement manual intervention and decrease the sepsis risk.

Automatic discovery of interpretable planning strategies

When making decisions, people often overlook critical information or are overly swayed by irrelevant information. A common approach to mitigate these biases is to provide decision-makers, especially professionals such as medical doctors, with decision aids, such as decision trees and flowcharts. Designing effective decision aids is a difficult problem. We propose that recently developed reinforcement learning methods for discovering clever heuristics for good decision-making can be partially leveraged to assist human experts in this design process. One of the biggest remaining obstacles to leveraging the aforementioned methods for improving human decision-making is that the policies they learn are opaque to people. To solve this problem, we introduce AI-Interpret: a general method for transforming idiosyncratic policies into simple and interpretable descriptions. Our algorithm combines recent advances in imitation learning and program induction with a new clustering method for identifying a large subset of demonstrations that can be accurately described by a simple, high-performing decision rule. We evaluate our new AI-Interpret algorithm and employ it to translate information-acquisition policies discovered through metalevel reinforcement learning. The results of three large behavioral experiments showed that providing the decision rules generated by AI-Interpret as flowcharts significantly improved people’s planning strategies and decisions across three different classes of sequential decision problems. Moreover, our fourth experiment revealed that this approach is significantly more effective at improving human decision-making than training people by giving them performance feedback. Finally, a series of ablation studies confirmed that our AI-Interpret algorithm was critical to the discovery of interpretable decision rules and that it is ready to be applied to other reinforcement learning problems. We conclude that the methods and findings presented in this article are an important step towards leveraging automatic strategy discovery to improve human decision-making. The code for our algorithm and the experiments is available at https://github.com/RationalityEnhancement/InterpretableStrategyDiscovery.

Separation of clods and stones from harvested potatoes using laser backscattering imaging technique

Traditionally, the separation of harvested potatoes from foreign materials, such as clods and stones, is performed manually with low efficiency and economic drawbacks. In this study, we proposed an automatic potato separation strategy that incorporates feature selection using laser backscattering imaging technology. Four sample mixtures of potatoes, clods, and stones from China were used for feature selection. Specifically, five optimal features were selected using combination of five filter methods and Borda count in the four groups. The classification model accounted for potato origin and variety was tested at different impurity rates. This approach presented high and promising separation results with an overall accuracy rate greater than 96%. The results demonstrated that the selected features were exceptionally powerful and effective for separating potatoes from non-potatoes of different origins and varieties under different impurity rates.

Generalized cross validation for \u2113p-\u2113q minimization

Discrete ill-posed inverse problems arise in various areas of science and engineering. The presence of noise in the data often makes it difficult to compute an accurate approximate solution. To reduce the sensitivity of the computed solution to the noise, one replaces the original problem by a nearby well-posed minimization problem, whose solution is less sensitive to the noise in the data than the solution of the original problem. This replacement is known as regularization. We consider the situation when the minimization problem consists of a fidelity term, that is defined in terms of a p-norm, and a regularization term, that is defined in terms of a q-norm. We allow 0 < p,q ≤ 2. The relative importance of the fidelity and regularization terms is determined by a regularization parameter. This paper develops an automatic strategy for determining the regularization parameter for these minimization problems. The proposed approach is based on a new application of generalized cross validation. Computed examples illustrate the performance of the method proposed.

Preliminary clinical experience of robot-assisted surgery in treatment with genioplasty

Genioplasty is the main way to treat diseases such as chin asymmetry, dysplasia and overdevelopment, which involve the three-dimensional direction abnormalities of the chin. Since this kind of surgery mainly uses intraoral incisions, the narrow surgical field of intraoral incisions and the surrounding important neurovascular tissues make it easy for complications, to occur during the osteotomy process, which results in greater surgical risks. The first craniofacial-plastic surgical robot (CPSR-I) system is developed to complete the precise positioning and improve the surgeon's force perception ability. The Kalman filtering method is adopted to reduce the interference of sensor signal noise. An adaptive fuzzy control system, which has strong robustness and adaptability to the environment, is designed to improve the stability of robot-assisted surgical operations. To solve the problem of the depth perception, we propose an automatic bone drilling control strategy that combines position and force conditions to ensure that the robot can automatically stop when the bone is penetrated. On the basis of model surgery and animal experiments, preliminary experiments were carried out clinically. Based on the early results of 6 patients, the robot-assisted approach appears to be a safe and effective strategy for genioplasty.

Technical Note: Patient-morphed mesh-type phantoms to support personalized nuclear medicine dosimetry - a proof of concept study.

Current standard practice for clinical radionuclide dosimetry utilizes reference phantoms, where defined organ dimensions represent population averages for a given sex and age. Greater phantom personalization would support more accurate dose estimations and personalized dosimetry. Tailoring phantoms is traditionally accomplished using operator-intensive organ-level segmentation of anatomic images. Modern mesh phantoms provide enhanced anatomical realism, which has motivated their integration within Monte Carlo codes. Here, we present an automatable strategy for generating patient-specific phantoms/dosimetry using intensity-based deformable image registration between mesh reference phantoms and patient CT images. This work demonstrates a proof-of-concept personalized dosimetry workflow, presented in comparison to the manual segmentation approach. A linear attenuation coefficient phantom was generated by resampling the PSRK-Man reference phantom onto a voxel grid and defining organ regions with corresponding Hounsfield unit (HU) reference values. The HU phantom was co-registered with a patient CT scan using Plastimatch B-spline deformable registration. In parallel, major organs were manually contoured to generate a "ground truth" patient-specific phantom for comparisons. Monte Carlo derived S-values, which support nuclear medicine dosimetry, were calculated using both approaches and compared. Application of the derived B-spline transform to the polygon vertices comprising the PSRK-Man yielded a deformed variant more closely matching the patient's body contour and most organ volumes as-evaluated by Hausdorff distance and Dice metrics. S-values computed for fluorine-18 for the deformed phantom using the Particle and Heavy Ion Transport code System showed improved agreement with those derived from the patient-specific analog. Deformable registration techniques can be used to create a personalized phantom and better support patient-specific dosimetry. This method is shown to be easier and faster than manual segmentation. Our study is limited to a proof-of-concept scope, but demonstrates that integration of personalized phantoms into clinical dosimetry workflows can reasonably be achieved when anatomical images (CT) are available.

An Adaptive Measurement System for the Simultaneous Evaluation of Frequency Shift and Series Resistance of QCM in Liquid.

In this paper, a novel measurement system based on Quartz Crystal Microbalances is presented. The proposed solution was conceived specifically to overcome the measurement problems related to Quartz Crystal Microbalance (QCM) applications in dielectric liquids where the Q-factor of the resonant system is severely reduced with respect to in-gas applications. The QCM is placed in a Meacham oscillator embedding an amplifier with adjustable gain, an automatic strategy for gain tuning allows for maintaining the oscillator frequency close to the series resonance frequency of the quartz, which is related in a simple way with the physical parameters of interest. The proposed system can be used to monitor simultaneously both the series resonant frequency and the equivalent electromechanical resistance of the quartz. The feasibility and the performance of the proposed method are proven by means of measurements obtained with a prototype based on a 10-MHz AT-cut quartz.

Quantifying Perception of Security Through Social Media and Its Relationship With Crime

The perception of security refers to the subjective evaluation of risks related to security events and the magnitude of their consequences. Negative perceptions have severe implications in society. These feelings are commonly quantified by citizen’s surveys, which are time-consuming and do not adapt well to the changing dynamics linked to security. Recently, Twitter social network emerged as an alternative to quantify some of these feelings dynamically. However, most of these approaches focused on counting the amount of content related to crime, which is only one of the relevant factors determining the perception of security. This work introduces a novel model for quantifying the Perception of Security on Twitter based on sentiment analysis and studies its relationship with actual crime. The model relies on an automatic strategy for filtering content related to security based on a support vector machine classifier and quantifying the sentiment in the posts using a multinomial naive Bayes classifier. This model studied more than 1.700.000 tweets in Bogota, collected during more than one year (March 18 of 2019 - April 28 of 2020). Results suggest that machine learning-based approaches may outperform previous security content filtering strategies by 21% in F1-score and provide quantifications on the sentiments of about 40% on accuracy. Trained models were used to provide daily estimates of perception of security, which are different from those captured by counting the number of posts related to security. Furthermore, the associations between perceptions and actual crimes exhibited monthly variations and showed that in some months, perceptions were more related to robbery, a crime with high incidence in the city. These results may help decision-makers devise strategies to reduce the impact of the negative perception on citizens.

Extraction of Family History Information From Clinical Notes: Deep Learning and Heuristics Approach

BackgroundElectronic health records store large amounts of patient clinical data. Despite efforts to structure patient data, clinical notes containing rich patient information remain stored as free text, greatly limiting its exploitation. This includes family history, which is highly relevant for applications such as diagnosis and prognosis.ObjectiveThis study aims to develop automatic strategies for annotating family history information in clinical notes, focusing not only on the extraction of relevant entities such as family members and disease mentions but also on the extraction of relations between the identified entities.MethodsThis study extends a previous contribution for the 2019 track on family history extraction from national natural language processing clinical challenges by improving a previously developed rule-based engine, using deep learning (DL) approaches for the extraction of entities from clinical notes, and combining both approaches in a hybrid end-to-end system capable of successfully extracting family member and observation entities and the relations between those entities. Furthermore, this study analyzes the impact of factors such as the use of external resources and different types of embeddings in the performance of DL models.ResultsThe approaches developed were evaluated in a first task regarding entity extraction and in a second task concerning relation extraction. The proposed DL approach improved observation extraction, obtaining F1 scores of 0.8688 and 0.7907 in the training and test sets, respectively. However, DL approaches have limitations in the extraction of family members. The rule-based engine was adjusted to have higher generalizing capability and achieved family member extraction F1 scores of 0.8823 and 0.8092 in the training and test sets, respectively. The resulting hybrid system obtained F1 scores of 0.8743 and 0.7979 in the training and test sets, respectively. For the second task, the original evaluator was adjusted to perform a more exact evaluation than the original one, and the hybrid system obtained F1 scores of 0.6480 and 0.5082 in the training and test sets, respectively.ConclusionsWe evaluated the impact of several factors on the performance of DL models, and we present an end-to-end system for extracting family history information from clinical notes, which can help in the structuring and reuse of this type of information. The final hybrid solution is provided in a publicly available code repository.