Trained Subjects Research Articles

Predicting Alzheimer’s disease progression in healthy and MCI subjects using multi‐modal deep learning approach

AbstractBackgroundAlzheimer’s disease (AD) is a complex disorder influenced by many factors, but it is unclear how each factor contributes to disease progression. An in‐depth examination of these factors may yield an accurate estimate of time‐to‐conversion to AD for patients at various disease stages. Recent advances in deep learning have enabled researchers to predict patient’s disease onset time by exploring the influencing factors in AD progression.MethodWe used 543 subjects with 63 features from 3 data modalities from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. The following modalities were used: 1) MRI, 2) genetic and 3) DTC (Demographic, cognitive Tests and Cerebrospinal fluid). The 21 most important features were automatically selected for the three modalities.We used a Deep Learning‐based survival analysis model that extends the classic Cox regression model to predict the subjects' disease onset time. Here we re‐define the non‐AD‐progression as “survivor”, and AD‐progression as “non‐survivor”. The subjects were divided into two groups: progressive subjects (non‐survivor), who were either healthy or diagnosed with Mild Cognitive Impairment (MCI) at initial clinical visit and later developed AD, and non‐progressive subjects ("survivor”), who were either healthy or MCI at initial visit but did not develop AD later. We used 10 random sub‐samples, selecting 80% of the subjects for training and 20% for testing each time; 20% of training data was used for internal validation.ResultFigure 1 shows the estimated survival rates over 10 years. Both groups had a high survival chance at the start. The progressive group’s survival chance dropped much faster and fell below 20% by the end of the period. The non‐progressive group’s survival chance remained around 50%.Feature importance analysis is displayed in Figure 2. Eight of the top ten most important features are from the cognitive tests, demonstrating their importance in survival analysis. Amygdala and Hippocampus regions, as well as age, are also notable features.ConclusionOur study demonstrated that using powerful predictive models on multi‐modal data can improve prediction of time‐to‐conversion. This not only leads to a better understanding of AD, but also provides essential tools for practitioners who wish to follow their patients' disease progression.

AI-Powered Noncontact In-Home Gait Monitoring and Activity Recognition System Based on mm-Wave FMCW Radar and Cloud Computing

In this work, we present a cloud-based system for non-contact, real-time recognition and monitoring of physical activities and walking periods within a domestic environment. The proposed system employs standalone Internet of Things (IoT)-based millimeter wave radar devices and deep learning models to enable autonomous, free-living activity recognition and gait analysis. To train deep learning models, we utilize range-Doppler maps generated from a dataset of real-life in-home activities. The performance of several deep learning models is evaluated based on accuracy and prediction time, with the gated recurrent network (GRU) model selected for real-time deployment due to its balance of speed and accuracy compared to 2D Convolutional Neural Network Long Short-Term Memory (2D-CNNLSTM) and Long Short-Term Memory (LSTM) models. The overall accuracy of the GRU model for classifying in-home physical activities of trained subjects is 93%, with 86% accuracy for a new subject. In addition to recognizing and differentiating various activities and walking periods, the system also records the subject’s activity level over time, washroom use frequency, sleep/sedentary/active/out-of-home durations, current state, and gait parameters. Importantly, the system maintains privacy by not requiring the subject to wear or carry any additional devices.

Teacher training, training teacher in charge and for change

This article brings bibliographical reflections on the subject of teacher training and the professionals who work in it, as a teacher and pedagogical coordinator, it aims to broaden the discussions on the subject, bringing the importance of continuing education in the professional development of the teacher and, concomitantly with it, the student learning in the school space.

A review of research on directional off-road path planning algorithms

Military orienteering can cultivate the skills of map reading, map use and military charting, and has gradually become a regular military training subject. A good track design plan is the basis for the smooth running of the competition. The current research on cross-country courses is mainly focused on single-track optimization, and there is very little research on multi-track planning, and design. However, in specific events, there are often multiple tracks running at the same time, and there are many restrictions, single track optimization design solutions are often not applicable. In this paper, we review multiple path planning algorithms and summarize the problems of these algorithms. On this basis, the advantages and disadvantages of static and dynamic path planning algorithms are analyzed, and future research trends and methods of path planning algorithms in orienteering are proposed.

Application of european society of cardiology guidelines 2020 on competitive sports eligibility in a large population of subjects with surgically corrected congenital heart disease

Abstract Funding Acknowledgements Type of funding sources: None. Background Congenital heart disease (CHD) affects approximately 0.5-1% of the general population. Thanks to the improvement of medical and surgical techniques, in developed countries 90% of individuals with CHD reach adulthood and, consequently, the demand for participation in sporting activities also increases. Recently, the European Society of Cardiology have given recommendations on participation in recreational and competitive sports (ESC 2020) based on a clinical and functional algorithm. Purpose To show the prevalence of the eligibility for competitive sports proposed by ESC 2020 in a real CHD population and the prevalence of real engagement in physical activity (both competitive and non-competitive). Also, to evaluate the difference in cardiorespiratory fitness (CRF) and the impact of training among different sport eligibility classes. Methods Retrospective study of 401 patients with surgically corrected CHD and followed up with cardiopulmonary exercise test (CPET) performed on treadmill or cycle ergometer until patients’ exhaustion. Subjects were divided in 4 classes according with ESC 2020 guidelines and corresponding to the eligibility for all competitive sport (A), all except endurance specialties (B), only for skill disciplines (C) or not eligible (D). Results Median (IQR) age was 18.0(13.0; 21.5) years; 266 patients (66.3%) were males. 34% were in class D, 13% in class B, 25% in class C and 28% in class A. Within the entire population, 179 subjects (44.6%) carried out structured exercise training, corresponding to 60.7% of subjects in class A, 45.1% of those in class B, 41.7% of those in class C and 41.1% of those in class D. VO2peak/Kg was similar between class A [39.2 (33.6-44.4) mL/Kg/min] and B [37.9 (30.6-42.6) mL/Kg/min], but significantly different between class A vs. class C [33.3 (28.3-41.4) mL/Kg/min p<.0001] and class D [32.2 (27.6-39.3) mL/Kg/min p<.0001]. Also, functional impairment, defined as <85% of the predicted for age and sex, resulted significantly higher in patients in the lower classes of sport eligibility (C and D) than in the higher ones, with a prevalence of impaired CRF of 22.5% in class A, 42.5% in class B, 55.1% in class C and 57.0% in class D (p <.0001). Paired comparison showed significant differences in CRF (VO2peak/Kg) between trained and untrained subjects both within and across ESC 2020 classes, while there was no significant difference between paired groups of different classes who presented the same level of physical activity (Figure 1). Conclusions Subjects in class A had better functional capacity than those in C and D classes, but exercise training (competitive or non-competitive) can have a significant impact on CRF level. Figure 1 shows that trained people in D class had higher CRF than untrained in A class, supporting that functional evaluation and an individualized exercise prescription may recruit functional "reserve" by improving the VO2peak as key prognostic component.

Aporte al ejercicio del derecho humano a la educación: propuestas formadoras de docentes, instituciones educativas y docentes

It is based on two central concepts to carry forward, the first in education and the second in human rights. The importance lies in the fact that each person has the right to receive quality education based on human rights, where each subject can enjoy a protected, inclusive and free life. It is essential to analyze the practice of education from a perspective that starts from human rights directly linked to social inclusion, in addition to the possibility of making proposals that help and frame this type of contribution of the exercise of human rights in education. Education based on human rights provides different spaces for action where a political-pedagogical activity is interacted that has the purpose of building in the classroom aspects that provide better inclusion and social equity for all. Education must be guaranteed by States because access to it is a right for all individuals and starting from an approach towards the exercise of a human right is to think that it can be an engine to train subjects in these bases and that truly any person can exercise it freely. Education has a fundamental role for development and changes in different societies, these changes have been made within and from human rights could not only be a challenge, but would bring immediate improvements in the quality of subjects individually and collectively.

Muscle Hypertrophy Responses to Changes in Training Volume: A Retrospective Analysis.

This study assessed associations between changes in the weekly number of sets performed and the percentage change in muscle thickness of the biceps (MTBB) and triceps (MTTB) brachii muscles. Through a retrospective analysis, sixty-eight resistance-trained subjects that participated in previous studies had their previous training volumes analyzed and compared to the volume imposed during each individual study. The relationship between variables was determined through Spearman correlation and a k-cluster analysis was performed to subdivide the participants into three groups and classified as:<0%; 0-50%, and >50% increase in the number of sets for both muscle groups. Moderate and weak correlations were observed between the alterations in training volume and changes in MTBB (rs=0.44, p=0.001) and MTTB (rs=0.35, p=0.002), respectively. A significant difference was noted between<0% to >50% for MTBB and MTTB (p=0.017; p=0.042, respectively), while no significant difference was observed between<0% to 0-50% and 0-50% to >50% (both p >0.05) for both muscle groups. In conclusion, muscle hypertrophy of the upper limbs is only weakly to moderately associated with changes in training volume of trained subjects.

Analysis of Spanish vocational radiographers' education through clinical training and perceptions of core subject teachers: A survey

IntroductionThis work presents a prospective analysis of the key aspects involved in the Spanish radiographer's perception of the weaknesses of the current educational curriculum in terms of teaching staff's qualifications and composition in the clinical training and core subjects. The goal is to show clinical training and professional's perception of the teaching quality and to characterise the weakness detected within the context of the European radiographer's academic system. MethodsThe perception of the quality of the training received by the professionals was collected through an anonymous survey. A total of 758 valid responses were received, and were analysed to three hypotheses: the variability of the teachers’ qualifications in the core subjects, the variability in terms of internships time carried out by the students, and assessment about teaching quality of the teachers. ResultsThe results prove there is a huge variability and little academic affinity of the teachers' degrees with the core subjects. On the other hand, the results shows there is a scarcity of clinical training hours in Spain, especially compared to European standards. It was demonstrated that teachers with a specific radiographer degree obtained the best scores. ConclusionIt is mandatory to adjust the criteria for selecting clinical imaging teachers to improve the teaching quality in Spain and increase the clinical training period of the Spanish radiographers to match their European counterparts. Implication for practiceImproving the training of Spanish radiographers will contribute to standardising the training quality of the whole European profession.

Resilient Coordinated Data-Driven Control of Multiple High-Speed Trains Under Fading Measurements and Denial-of-Service Attacks

The paper investigates the resilient coordinated model free adaptive control (MFAC) problem of multiple high-speed trains (MHSTs) subject to fading measurements and denial-of-service (DoS) attacks, which belongs to a data-driven approach. Firstly, the dynamic model of MHSTs with unknown nonlinear dynamics is linearized into a data-dependent one. Secondly, the faded channels are formulated by the Rice fading model with known statistical property, and the attack strategies are assumed to obey the Bernoulli distributions. Then, the distributed MFAC method is proposed for MHSTs with fixed topology, and the analysis is carried out by using the graph theory. Furthermore, the research is extended to the design of the compensation scheme for fading data and to the MHSTs with time-varying topologies. Finally, the effectiveness of the data-driven controller is studied by a numerical simulation.

Assessing the perception of face pareidolia in children (Homo sapiens), rhesus monkeys (Macaca mulatta), and capuchin monkeys (Sapajus apella).

Face pareidolia is the misperception of a face in an inanimate object and is a common feature of the face detection system in humans. Whereas there are many similarities in how humans and nonhuman animals such as monkeys perceive and respond to faces, it is still unclear whether other species also perceive certain nonface stimuli as faces. We presented a novel computerized task to capuchin monkeys (Sapajus apella), rhesus monkeys (Macaca mulatta), and preschool-aged children (Homo sapiens). This task trained subjects to choose faces over nonface images, and then presented pareidolia images with nonface images. All species selected faces most often on trials that included face images. However, only children selected pareidolia images at levels above chance. These results indicate that while children report perceiving face pareidolia, monkeys do not. These species differences could be due to human-unique experiences that result in an increased aptitude for anthropomorphizing objects with face-like patterns. It could also be due to monkeys showing a greater reliance on stimulus features rather than global, holistically organized cues that faces provide. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Always Be Prepared

Always Be Prepared

C-3 | A Machine Learning Algorithm to Detect Right Heart Failure in Cardiogenic Shock when Using Impella

Right heart failure (RHF) is common amongst patients in cardiogenic shock (CS) and best diagnosed with use of a pulmonary artery catheter (PAC). Unfortunately use of PACs remains low (<20%) even in cases of CS. Early identification and treatment of RVF may improve outcomes for such patients. We propose a machine learning algorithm to detect RVF in patients presenting with cardiogenic shock who are treated with an Impella CP. We leveraged clinical data collected in the National Cardiogenic Shock Initiative (NCSI) to develop and train a RHF Indicator model: RHFNet, which is a dense neural network model that can infer the likelihood of RHF from Impella CP data (Figure). The model was developed using 129 NCSI patients (N=129 development subjects: 49 (32%) RHF, 80 (68%) No RHF) on the data from first 24 hours of Impella CP support. For validation, 84 independent cases were identified from the Abiomed IQ database that had concurrent CVP and PAPI measurements. These cases were assigned to a cohort based on their CVP and PAPI values to create a validation set. RHFNet was trained on a 66% random sample of development subjects (N=86 training subjects) and tested on the remaining 34% subjects (N=43 test subjects) to establish the viability of the solution. The model achieved a mean AUC score of 0.80 on the test subjects. On independent IQ cases (N=84 validation subjects), RHFNet achieved a similar AUC score of 0.78. RHF can be identified using only signals from a left-sided MCS device (Impella CP).

Dynamics of special physical training of military officers of the technical supply of the armies

Achieving high results in military service, as in any activity, depends to a certain extent on the features of the established training and education process, which must be developed and improved taking into account reasonable methods of influence, which in turn correspond to the capabilities and needs of each serviceman individually. The effectiveness of the servicemen's work is determined and depends on his level of professional training, namely full and high-quality mastery of the subjects of military-professional training.&#x0D; Physical training in the Armed Forces plays an important role and ensures the physical readiness of servicemen for combat activities, the performance of official tasks as assigned, contributes to solving other tasks, their training and education. It is during physical training that basic physical qualities (strength, speed, endurance) are formed and developed, and special qualities and applied skills of a military nature are improved, which are necessary throughout the entire military service of a serviceman, and certain moral and psychological qualities are also developed, which contributes to the formation the warriors have a professional outlook and professional self-awareness of the individual.

Learning design for couple prenatal education program; birthing class

&lt;p&gt;&lt;strong&gt;&lt;em&gt;Background&lt;/em&gt;&lt;/strong&gt;&lt;em&gt;: The involvement of birth attendants has been shown to have a positive impact on pregnancy outcomes. Unprepared fathers tend to have negative experiences during childbirth. Mothers and partners need to be prepared since pregnancy to face childbirth and have a satisfying birth experience. It is necessary to arrange a learning program for childbirth classes not only for pregnant woman but also their partners.&lt;/em&gt;&lt;/p&gt;&lt;p&gt;&lt;strong&gt;&lt;em&gt;Objectives:&lt;/em&gt;&lt;/strong&gt;&lt;em&gt; The purpose of this study is to develop a birthing class learning design for the prenatal couple education program.&lt;/em&gt;&lt;/p&gt;&lt;p&gt;&lt;strong&gt;&lt;em&gt;Methods&lt;/em&gt;&lt;/strong&gt;&lt;em&gt;: This research design is research and development (R&amp;amp;D) level 1. The product to be developed is a birthing class learning design for prenatal couple education program. The stages of the research are 1) analysis of gaps and needs of the developed program; 2) Drafting; 3) internal validation of the design &lt;/em&gt;&lt;/p&gt;&lt;p&gt;&lt;strong&gt;&lt;em&gt;Results&lt;/em&gt;&lt;/strong&gt;&lt;em&gt;: This study found that the needs of mother and their partners for birthing classes are high. The material most needed in birthing class is husband’s role in delivery process. In the management of birthing class, mother and partner want a balance practicum and theory method, the duration of the class is 90 minutes, with the number of participants less than 5 couples and the use of learning modules. The birthing class learning design for prenatal couple education program is structured in form of training program curriculum that consist 1) competencies; 2) learning outcome; 3)study materials; 4) training subjects; 5) training course loads; 6) learning methods and 7) assessments. &lt;/em&gt;&lt;/p&gt;&lt;strong&gt;&lt;em&gt;Conclusions&lt;/em&gt;&lt;/strong&gt;&lt;em&gt; The birthing class learning design for prenatal couple education program is needed by pregnant women and their partners. The design can be used for preparing the birthing class model.&lt;/em&gt;

Hearing Range Analysis of Rattus Argentiventer the Paddy Field Pest

Rice is the second most-grown cereal crop and one of the important staple foods in Malaysia. Seasonal rodent attacks from a family of Rattus argentiventer have shortened rice production by up to 80%. Yet, the acoustical method in controlling rodent populations is one of the topics that have not been properly explored. Under the field study radius, this study aims to analyse the hearing range of R. argentiventer under a controlled environment. Knowing a rodent’s ability to hear is essential for evaluating whether human activities, especially in terms of noise pollution, have an impact on hearing and consequently on rodent behaviour. Thirteen subjects were randomly cage-caught in the paddy field and were tested in a reward-based go/no-go procedure. Fully trained subjects were observed after acoustically exposing the subjects to a 10 kHz of pure tone less than 60 dB SPL for 25 sessions. Hearing range analysis commenced by revealing a pure tone to the fully trained subjects at low (1 kHz – 5 kHz), median (10 kHz – 40 kHz) and high (45 kHz – 80 kHz) frequencies. The results showed the subjects have sensitive frequencies at 5 kHz for low frequency, the median frequency at 25 kHz and high frequency at 45 kHz. From the sensitivity hearing results, it can be indicated that the hearing range of R. argentiventer is between 1 kHz to 55 kHz at sound levels of 60 dB. The results responded to the lack of timely response which is within an average of 5 sec and the minimum licking period at an average of 13 sec as the inaudible cut-off hearing range.

The impact of a story on learning ketone body metabolism.

Biochemistry is a core subject in the cross-disciplinary training on Biotechnology engineering courses. Metabolic pathways teaching has traditionally integrated hands-on laboratory experiences and traditional lectures, which detail a large number of reactions at a molecular level, their enzymes and regulation. The current scenario of Covid-19 outbreak have motivated the development of complementary tools that expand the horizon of metabolism teaching. In this study, we employed a story-based methodology to strengthen the metabolic pathways learning and to measure students' perception. Specifically, a peer-reviewed tale describing the ketone body metabolism was used during five semesters as a didactic strategy to teach this biochemical process. A questionnaire assessed the students' understanding and acceptance of the methodology (n = 83). Our findings showed that a high proportion of students (83.13%) were able to relate the story to the topics studied in the classroom (ketogenesis and ketolysis). On the other hand, they were satisfied and suggested that such methodology is effective and fun. In summary, most of the survey responses related to acceptance of story-based strategy ranged from 72% to 97%. Collectively, these results indicated that the story is appropriate to decomplex pathways, becoming a simple tool for driving motivation, learning and engagement of students. The narrative represents a bridge to connect the intriguing series of chemical reactions involved in the anabolism and degradation of 3-hydroxybutyrate (3-OHB), acetoacetate, and acetone with previously learned knowledge, emotions, and key concepts. In conclusion, the tale was useful to decode ketone body-related pathways and making metabolism learning more interesting and easier.

Impact of high-intensity interval training on tendon related gene expression in rat Achilles tendon

Immobilization or aging associated with limited physical activity can lead to the functional deterioration of tendons, which has become an important public health concern. Therefore, growing research is focused on the effect of exercise training on preserving tendon function. Exercise training subjects muscles and tendons to repeated mechanical stress, and in vitro studies have revealed that repetitive mechanical loading stimulates tendon cell responses to changes in the extracellular matrix and functional properties of the tendon. However, although several types of exercise training have been shown to be effective in preserving tendon function, no studies have investigated the impact of high-intensity interval training (HIIT), which involves composing short bouts of exercise with high-power output. Here, we determined whether the HIIT program enhanced tenogenic progressions by measuring the mRNA expression in rat Achilles tendons. Sixteen rats were randomly assigned into either a sedentary control group (Con, n = 8) or an HIIT group (n = 8). Rats in the HIIT group performed the program with treadmill running; the training volume was incremental (running speed, number of sets, and inclination), and training was conducted 5 days per week for 9 weeks. The rats in the HIIT group exhibited a marked decrease in the body weight and different types of fat weights, and a marked increase in different types of muscle weights. Real-time reverse transcription polymerase chain reaction analysis revealed that mRNA expressions of tendon-related genes Tnxb, Opn, and Tgfb1 were upregulated in the HIIT group compared to that in the Con group. Cross-links in mRNA expressions of collagen-related Dcn and Fmod in the HIIT group tended to be higher than in those Con group. These results suggest that HIIT induces initiation of tenogenic progression and stimulation of cross-link formation between collagen fibrils in rat Achilles tendons.

경비원 전문성 향상을 위한 자격제도 도입방안

The qualitative growth as well as the quantitative growth of the Korean private security industry is that the private security is in charge of the function of substantially supplementing the police force in protecting the life, body and property of the people. And that's why it's necessary to study. This study, which presented a plan to introduce a qualification system to improve the professionalism of security guards, drew the following conclusions. First, the security work must be clearly classified and the qualification certificates must be subdivided according to the field and scope of the security work. Second, it is necessary to distinguish between qualifications for private security guard examination. Third, in the evaluation method, not only the department test but also the practical test should be conducted in a complex way. Fourth, evaluation should be conducted through the connection between department and practical course. Currently, the security industry is complexly intertwined with various issues such as scope of work, education, management, and supervision, so a complex study is needed rather than a single study. - It is necessary to make a leap and develop into one specialized field through various research and support, such as deriving job-related education and training subjects, roles and changes of management and supervision institutions, and utilization of security-related professionals. In this case, safety problems such as congestion and terrorism due to the participation of citizens and tourists are emerging. The causes of these problems have also changed as society has diversified, and safety measures for large-scale events involving citizens must also change according to the changing social threat factors.

Central nodes of canine functional brain networks are concentrated in the cingulate gyrus

Compared to the field of human fMRI, knowledge about functional networks in dogs is scarce. In this paper, we present the first anatomically-defined ROI (region of interest) based functional network map of the companion dog brain. We scanned 33 awake dogs in a “task-free condition”. Our trained subjects, similarly to humans, remain willingly motionless during scanning. Our goal is to provide a reference map with a current best estimate for the organisation of the cerebral cortex as measured by functional connectivity. The findings extend a previous spatial ICA (independent component analysis) study (Szabo et al. in Sci Rep 9(1):1.25. https://doi.org/10.1038/s41598-019-51752-2, 2019), with the current study including (1) more subjects and (2) improved scanning protocol to avoid asymmetric lateral distortions. In dogs, similarly to humans (Sacca et al. in J Neurosci Methods. https://doi.org/10.1016/j.jneumeth.2021.109084, 2021), ageing resulted in increasing framewise displacement (i.e. head motion) in the scanner. Despite the inherently different approaches between model-free ICA and model-based ROI, the resulting functional networks show a remarkable similarity. However, in the present study, we did not detect a designated auditory network. Instead, we identified two highly connected, lateralised multi-region networks extending to non-homotropic regions (Sylvian L, Sylvian R), including the respective auditory regions, together with the associative and sensorimotor cortices and the insular cortex. The attention and control networks were not split into two fully separated, dedicated networks. Overall, in dogs, fronto-parietal networks and hubs were less dominant than in humans, with the cingulate gyrus playing a central role. The current manuscript provides the first attempt to map whole-brain functional networks in dogs via a model-based approach.

MASS: A Multisource Domain Adaptation Network for Cross-Subject Touch Gesture Recognition

Touch gesture recognition (TGR) plays a pivotal role in many applications, such as socially assistive robots and embodied telecommunication. However, one obstacle to practicality of existing TGR methods is the individual disparities across subjects. Moreover, a deep neural network trained with multiple existing subjects can easily lead to overfitting for a new subject. Hence, how to mitigate the discrepancies between the new and existing subjects and establish a generalized network for TGR is a significant task to realize reliable human–robot tactile interaction. In this article, a novel framework for Multisource domain Adaptation via Shared-Specific feature projection (MASS) is proposed, which incorporates intradomain discriminant, multidomain discriminant, and cross-domain consistency into a deep learning network for cross-subject TGR. Specifically, the MASS method first extracts the shared features in the common feature space of training subjects, with which a domain-general classifier is built. Then, the specific features of each pair of training and testing subjects are mapped and aligned in their common feature space, and multiple domain-specific classifiers are trained with the specific features. Finally, the domain-general classifier and domain-specific classifiers are ensembled to predict the label for the touch samples of a new subject. Experimental results performed on two datasets show that our proposed MASS method achieves remarkable results for cross-subject TGR. The code of MASS is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/AI-touch/MASS</uri> .