Sequential Training Research Articles

Prospective validation of DACH2 as a novel biomarker for prediction of metastasis and prognosis in muscle-invasive urothelial carcinoma of the bladder

Metastasis is the main cause of death from muscle-invasive urothelial carcinoma of the bladder (UCB), and the metastatic potential of tumors is often unpredictable. The role of Dachshund homolog 2 gene (DACH2) in tumorigenesis remains unexplored. We aimed to investigate whether DACH2 can be used as a biomarker to predict metastasis and prognosis of muscle-invasive UCB in a sequential training and validation fashion. For the training set (n = 40), compared with UCB patients without lymph node (LN) metastasis, both DACH2 protein and mRNA expression were greatly increased in case-matched patients with LN metastasis. For the independent validation set (n = 243), patients with primary UCB that did not express DACH2 had a longer metastasis-free survival (MFS) and overall survival (OS) than did those with tumors expressing DACH2 (5-year MFS: 88% [95% CI 80–96] versus 19% [95% CI 7–31], p < 0.001; 5-year OS: 93% [95% CI 87–99] versus 37% [95% CI 23–51], p < 0.001). Multivariable analysis of DACH2 status showed hazard ratios of 7.34 (95% CI 3.15-11.87, p < 0.001) for MFS and 3.96 (95% CI 2.04-7.16, p < 0.001) for OS which were much higher than hazard ratios associated with other independent risk factors. Collectively, DACH2 is an independent prognostic marker that can be used at initial diagnosis of UCB to identify patients who have a high potential to develop metastasis.

Clustering Fragments Metagenome Using Self-Organizing Map

Metagenome is a combination of several microorganisms collected from the environment. In metagenome analysis, it is required binning for grouping metagenome fragment yielded by sequencer. This research used the composition approach for conducting metagenome fragment binning. In this approach, binning could be implemented using unsupervised or supervised learning. We used Self-Organizing Map (SOM) for conducting binning used on unsupervised learning. We compared two techniques of training in SOM, namely sequential training and batch training for finding the best techniques. The results showed that the batch training could obtain 3.8% error valued on the map of [10 15]. This error value is smaller than that of sequential training. DOI: http://dx.doi.org/10.11591/telkomnika.v13i3.7115

Colonoscopy training with a computer–animal–clinic sequential training model: a randomized controlled trial

Objective To explore the value of computer–animal–clinic sequential training model in colonoscopy training. Methods Sixty–five gastroenterologists without any experience in colonoscopy were randomly assigned to 2 groups to accept a two–week' endoscopy training with computer simulator or with computer–animal–clinic sequential training. Both groups underwent at least 10 hours of training with a computer–based simulator, and sequential training group also underwent animal simulator training for 2 days in first week, then both groups underwent clinical endoscopy training in the second week. Performance parameters including cecal intubation, procedure duration and degree of comfort were evaluated. Results The training group had higher cecal intubation rate (46.28% VS 35.79%; P<0.01) and shorter procedure duration (9.05±2.12 min VS 11.15±3.12 min; P<0.05) and less degree of comfort (5.18±1.41 VS 6.78±2.15; P<0.05). The sequential training group was much better in performance than the other group in colonoscopy training. Conclusion Computer–animal–clinic sequential training model is effective in providing trainees with colonoscopy skills in improving the success rate, shortening the teaching times and lessening the uncomfortable of patients. Key words: Endoscopy; Education, professional, retraining; Computer simulation; Animal experimentation

Sequencing bilateral and unilateral task-oriented training versus task oriented training alone to improve arm function in individuals with chronic stroke.

BackgroundRecovering useful hand function after stroke is a major scientific challenge for patients with limited motor recovery. We hypothesized that sequential training beginning with proximal bilateral followed by unilateral task oriented training is superior to time-matched unilateral training alone. Proximal bilateral training could optimally prepare the motor system to respond to the more challenging task-oriented training.MethodsParticipants: Twenty-six participants with moderate severity hemiparesis Intervention: Participants received either 6-weeks of bilateral proximal training followed sequentially by 6-weeks unilateral task-oriented training (COMBO) or 12-weeks of unilateral task-oriented training alone (SAEBO). A subset of 8 COMB0 and 9 SAEBO participants underwent three functional magnetic resonance imaging (fMRI) scans of hand and elbow movement every 6 weeks. Main Outcome Measures: Fugl-Meyer Upper extremity scale, Modified Wolf Motor Function Test, University of Maryland Arm Questionnaire for Stroke, Motor cortex activation (fMRI).ResultsThe COMBO group demonstrated significantly greater gains between baseline and 12-weeks over all outcome measures (p = .018 based on a MANOVA test) and specifically in the Modified Wolf Motor Function test (time). Both groups demonstrated within-group gains on the Fugl-Meyer Upper Extremity test (impairment) and University of Maryland Arm Questionnaire for Stroke (functional use). fMRI subset analyses showed motor cortex (primary and premotor) activation during hand movement was significantly increased by sequential combination training but not by task-oriented training alone.ConclusionsSequentially combining a proximal bilateral before a unilateral task-oriented training may be an effective way to facilitate gains in arm and hand function in those with moderate to severe paresis post-stroke compared to unilateral task oriented training alone.

Inactivation of the anterior cingulate reveals enhanced reliance on cortical networks for remote spatial memory retrieval after sequential memory processing.

One system consolidation model suggests that as time passes, ensembles of cortical neurons form strong connections to represent remote memories. In this model, the anterior cingulate cortex (ACC) serves as a cortical region that represents remote memories. However, there is debate as to whether remote spatial memories go through this systems consolidation process and come to rely on the ACC. The present experiment examined whether increasing the processing demand on the hippocampus, by sequential training on two spatial tasks, would more fully engage the ACC during retrieval of a remote spatial memory. In this scenario, inactivation of the ACC at a remote time point was hypothesized to produce a severe memory deficit if rats had been trained on two, sequential spatial tasks. Rats were trained on a water maze (WM) task only or a WM task followed by a radial arm maze task. A WM probe test was given recently or remotely to all rats. Prior to the probe test, rats received an injection of saline or muscimol into the ACC. A subtle deficit in probe performance was found at the remote time point in the group trained on only one spatial task and treated with muscimol. In the group trained on two spatial tasks and treated with muscimol, a subtle deficit in probe performance was noted at the recent time point and a substantial deficit in probe performance was observed at the remote time point. c-Fos labeling in the hippocampus revealed more labeling in the CA1 region in all remotely tested groups than recently tested groups. Findings suggest that spatial remote memories come to rely more fully on the ACC when hippocampal processing requirements are increased. Results also suggest continued involvement of the hippocampus in spatial memory retrieval along with a progressive strengthening of cortical connections as time progresses.

Sequential Adaptive RBF-Fuzzy Variable Structure Control Applied to Robotics Systems

In this paper, we present a combination of sequential trained radial basis function networks and fuzzy techniques to enhance the variable structure controllers dedicated to robotics systems. In this aim, four RBFs networks were used to estimate the model based part parameters (Inertia, Centrifugal and Coriolis, Gravity and Friction matrices) of a variable structure controller so to respond to model variation and disturbances, a sequential online training algorithm based on Growing-Pruning GAP strategy and Kalman filter was implemented. To eliminate the chattering effect, the corrective control of the VS control was computed by a fuzzy controller. Simulations are carried out to control three degrees of freedom SCARA robot manipulator where the obtained results show good disturbance rejection and chattering elimination.

A new minimal training sample scheme for intrinsic Bayes factors in censored data

The problem of covariate selection for regression models with right censored data is considered. It is approached from a default Bayesian point of view with Bayes factors (BFs) and in particular with Intrinsic BF (IBF) that depends on the minimal training samples (MTSs). In the presence of censored data, the number of possible MTSs increases, due to the fact that uncensored data, relevant for training the improper prior into a proper posterior, must be combined with censored data. For this purpose, the sequential minimal training sample scheme (SMTS) accounts for such requirements but generally leads to IBF correction factors that do not have an analytical form and thus require numerical approximation. In order to obtain an analytical expression of the correction terms, a different TS scheme is introduced based on the Kaplan–Meier (KM) estimator, termed the KM minimal training sample scheme. This new tool works extremely well in the analyzed simulation setting and also in the applications; it produces results which are similar, if not better, than the IBF calculated using MTSs. The resulting new IBF, being based on analytical expressions, is much faster to compute. Evidence of these results comes from a large simulation study, theoretical arguments, and an application to a real data set.

ANN based self tuned PID like adaptive controller design for high performance PMSM position control

Proportional-integral-derivative (PID) being the most simple and the widely deployed controller in the industrial drives is not quite amenable to the solution for high performance drives as these drives are subjected to the parametric uncertainty, unmodeled dynamics and variable load conditions during operation. In order to expand the robustness and adaptive capabilities of conventional PID controller, a neural network based PID (NNPID) like controller which is tuned when the controller is operating in an on line mode for high performance permanent magnet synchronous motor (PMSM) position control is proposed in this paper. The NN based PID like controller is composed of a mixed locally recurrent neural network and contains at most three hidden nodes which form a PID like structure. A novel training algorithm for the PID controller gain initialization based upon the minimum norm least square solution is proposed. An on line sequential training algorithm based on recursive least square is then derived to update controller gains in an on line manner. The proposed controller is not only easy to implement but also requires least number of parameters to be tuned prior to the implementation. The performance of the proposed controller is evaluated in the presence of parametric uncertainties and load disturbances also the result outcomes are compared with the conventional PID controller, optimized using Cuckoo search based optimization method.

The Effects of High Intensity Short Rest Resistance Exercise on Muscle Damage Markers in Men and Women

Within and between sexes, universal load prescription (as assigned in extreme conditioning programs) creates extreme ranges in individual training intensities. Exercise intensity has been proposed to be the main factor determining the degree of muscle damage. Thus, the purpose of this study was to examine markers of muscle damage in resistance-trained men (n = 9) and women (n = 9) from a high intensity (HI) short rest (SR) (HI/SR) resistance exercise protocol. The HI/SR consisted of a descending pyramid scheme starting at 10 repetitions, decreasing 1 repetition per set for the back squat, bench press, and deadlift, as fast as possible. Blood was drawn pre-exercise (pre), immediately postexercise (IP), 15 minutes postexercise (+15), 60 minutes postexercise (+60), and 24 hours postexercise (+24). Women demonstrated significant increases in interleukin 6 (IL-6; IP), creatine kinase (CK; +24), myoglobin (IP, +15, +60), and a greater relative increase when compared with men (+15, +60). Men demonstrated significant increases in myoglobin (IP, +15, +60, +24), IL-6 (IP, +15), CK (IP, +60, +24), and testosterone (IP, +15). There were significant sex interactions observed in CK (IP, +60, +24) and testosterone (IP, +15, +60, +24). Women completed the protocol faster (women: 34:04 ± 9:40 minutes, men: 39:22 ± 14:43 minutes), and at a slightly higher intensity (women: 70.1 ± 3.5%, men 68.8 ± 3.1%); however, men performed significantly more work (men: 14384.6 ± 1854.5 kg, women: 8774.7 ± 1612.7 kg). Overall, women demonstrated a faster inflammatory response with increased acute damage, whereas men demonstrated a greater prolonged damage response. Therefore, strength and conditioning professionals need to be aware of the level of stress imposed on individuals when creating such volitional high intensity metabolic type workouts and allow for adequate progression and recovery from such workouts.

Effects of concurrent and sequential training of maximum strength and oxygen endurance on certain components of the body composition

The study aimed at being acquainted with the effects of concurrent and sequential training of maximum strength and oxygen endurance on certain components of the body composition in addition to the differences between those two types of the training at the post test for the same variables. The experimental method was used. The study conducted on a sample of 20 subjects from the "Athletic Qaraqush Club". The sample divided into two groups with 10 subjects each. The homogeneity and equality were achieved in variables of age, high, weight and certain components of the body composition. The procedure of the study included determining the maximum intensities of the endurance and external weights for setting the intensities that will be used in the trainings. Two pilot studies were conducted. The first one was for the adopted exercise in the study and the second one was for conducting a training session. After that, the final field procedures were applied which started at pre tests of the body composition components followed by applying the two curricula of the concurrent and sequential trainings which lasted 9 weeks with 3 meso cycles. Each meso cycle included 3 micro cycles with (2:1) workload wave and with 3 training sessions per a week for the concurrent training and 6 training sessions per a week for the sequential one. Finally, the posts tests were conducted using the same procedure for the pre tests. After using the SPSS package to process the data, the conclusions were as the following:  The first experimental group, which applied the concurrent training, significantly decreased the fat weight and significantly increased the body weight.  The second experimental group, which applied the sequential training, significantly decreased the fat weight and significantly increased the free fat weight, body water and body weight.  There were significant differences between the two groups of the concurrent and sequential training in the free fat weight, body water and body weight so that they were tended to increase with the sequential training and decreased with the concurrent one. The most important recommendations of the study were to emphasize on coaches to use the concurrent training for improving the maximal strength, oxygen endurance, fat weight reduction, free fat weight, body water and body weight; and to emphasize on coaches to use the sequential training for reducing the fat weight in addition to raise the free fat weight, body water and body weight.

Acquisition of response thresholds for timed performance is regulated by a calcium‐responsive transcription factor, CaRF

Interval timing within the seconds-to-minutes range involves the interaction of the prefrontal cortex and basal ganglia via dopaminergic-glutamatergic pathways. Because the secreted protein brain-derived neurotrophic factor (BDNF) is able to modulate dopamine release as well as glutamatergic activity, we hypothesized that BDNF may be important for these timing mechanisms. Recently, the calcium-responsive transcription factor (CaRF) was identified as an important modulator of BDNF expression in the cerebral cortex. In this study, a strain of Carf knockout mice was evaluated for their ability to acquire the 'Start' and 'Stop' response thresholds under sequential and simultaneous training conditions, using multiple (15-second and 45-second) or single (30-second) target durations in the peak-interval procedure. Both Carf(+/-) and Carf(-/-) mice were impaired in their ability to acquire timed response thresholds relative to Carf(+/+) mice. Additionally, control mice given microinjections of BDNF antisense oligodeoxynucleotide to inhibit protein expression in the prefrontal cortex showed timing impairments during acquisition similar to Carf mice. Together, these results suggest that the inhibitory processes required to update response thresholds and exert temporal control of behavior during acquisition may be dependent on CaRF regulation of genes including Bdnf in cortico-striatal circuits.

A Revolution That Will Transform How We Live, Work, and Think: An Interview with the Authors of Big Data.

Big DataVol. 1, No. 2 InterviewsA Revolution That Will Transform How We Live, Work, and Think: An Interview with the Authors of Big DataEdd DumbillEdd DumbillSearch for more papers by this authorPublished Online:18 Jun 2013https://doi.org/10.1089/big.2013.0016AboutSectionsView articleView Full TextPDF/EPUB ToolsPermissionsDownload CitationsTrack CitationsAdd to favorites Back To Publication ShareShare onFacebookTwitterLinked InRedditEmail View articleFiguresReferencesRelatedDetailsCited ByRandom Forest Algorithm Based on Linear Privacy Budget AllocationJournal of Database Management, Vol. 33, No. 2Feature selection based on non-negative spectral feature learning and adaptive rank constraintKnowledge-Based Systems, Vol. 236Adaptive Rotation Forests: Decision Tree Ensembles for Sequential LearningSmoothing Adversarial Training for GNNIEEE Transactions on Computational Social Systems, Vol. 8, No. 3An optimal Best-Worst prioritization method under a 2-tuple linguistic environment in decision makingComputers & Industrial Engineering, Vol. 155Detection of Sociolinguistic Features in Digital Social Networks for the Detection of Communities26 January 2021 | Cognitive Computation, Vol. 13, No. 2Applying Advanced Technologies in Construction and Demolition Waste Management: A Review8 June 2021An Overview on CryptDb and Word2vec Approaches1 October 2020 | Advances in Science, Technology and Engineering Systems Journal, Vol. 5, No. 5Research trend of the application of information technologies in construction and demolition waste managementJournal of Cleaner Production, Vol. 263Stock Market Temporal Complex Networks Construction, Robustness Analysis, and Systematic Risk Identification: A Case of CSI 300 IndexComplexity, Vol. 2020A Review for Weighted MinHash AlgorithmsIEEE Transactions on Knowledge and Data EngineeringQuality Driven Judicial Data GovernanceResearch on big data feature analysis based on kernel discriminant analysis and neural network27 February 2018 | Multimedia Tools and Applications, Vol. 99Based on Hadoop's tech big data combination and mining technology frameworkImproved Consistent Weighted Sampling RevisitedIEEE Transactions on Knowledge and Data EngineeringHybrid datafication of maintenance logs from AI-assisted human tagsA Rapid Incremental Frequent Pattern Mining Algorithm for Uncertain DataBig Data and cloud computing: innovation opportunities and challenges3 November 2016 | International Journal of Digital Earth, Vol. 10, No. 1 Volume 1Issue 2Jun 2013 InformationCopyright 2013, Mary Ann Liebert, Inc.To cite this article:Edd Dumbill.A Revolution That Will Transform How We Live, Work, and Think: An Interview with the Authors of Big Data.Big Data.Jun 2013.73-77.http://doi.org/10.1089/big.2013.0016Published in Volume: 1 Issue 2: June 18, 2013PDF download

Can captive-bred American bullfrogs learn to avoid a model avian predator?

Animals that are isolated from their natural predators may lose the ability to express antipredator behavior. The aim of this study was to test whether it would be possible to train captive-bred North American bullfrogs (Rana catesbeiana) to avoid a model avian predator and to measure their behavioral responses to this predator. We used 18 animals divided into two groups, trained (N = 6) and control (N = 12); these individuals were from a line that has been bred in captivity for at least 20 generations without predator contact. The trained group was exposed, individually, to 20 sequential antipredator training sessions, during which the presence of a model avian predator was paired with an aversive stimulus. The control group was exposed, individually, once only to the same model avian predator, but without the association of the aversive stimulus. Both groups were observed for 10 min after the presentation of the avian predator, during which their behavior was recorded using instantaneous recording of behavior. The results showed that, after only two training sessions, the trained bullfrogs started to express proper antipredator behavior: diving to the bottom of the tank, lying still, and eventually, learning to cover themselves in substrate (mud). However, continued training sessions provoked a varied response to antipredator training, which was undesirable as the effectiveness of their antipredator response varied. This study has shown that captive-bred bullfrogs were capable of learning antipredator behavior, despite having been bred for many generations in captivity with no predator contact.

Sequential Stimulus Pairing Procedure for the Students with Intellectual Disabilities

For most of students with autism spectrum disorders (ASD) and Williams syndrome (WS), a rare genetic neurodevelopmental disorder, are known to have intellectual disabilities (ID). Students with ID often show the difficulties in reading. Especially, they are difficult to acquire the equivalence relations among pictures, written letters, and sounds and to have fluent eye movement during reading. Previous research suggested that a student with autism acquired Kanji reading skills by using stimulus pairing training. However, for acquiring word reading skills, new training which facilitates the fluent eye movement is necessary and we developed sequential stimulus pairing training. In the present study, we examined the acquisition of word reading skills through sequential stimulus pairing training for three students with ID who were also diagnosed as WS and three students with ID who were not diagnosed with WS. In a trial, each letters, the word, spoken sound, and picture were presented sequentially. With 6 students, result indicated that they could acquire the word reading skills, and also showed the improvement of their eye movement in reading. The result suggested sequential stimulus pairing training is effective to acquire both equivalence relations and fluent eye movement for wide range of students with ID.

Gastric endoscopic submucosal dissection: From animal model to patient

To assess whether the use of porcine models is useful for learning endoscopic submucosal dissection (ESD), thus contributing to its subsequent application in human patients. This study/learning process was carried out in 3 phases: Phase I: Ex vivo animal; Phase II: In vivo animal; Phase III: Humans. One endoscopist performed 30 gastric ESDs in porcine models, and later 5 gastric ESDs in 5 patients. The ESD was done following the method practiced at the National Cancer Center in Tokyo, Japan. Technical aspects, size, time and speed of ESD, as well as complications were registered. In patients, their clinical, endoscopic and histologic evolution was additionally added. Thirty en bloc ESDs were carried out in animal models. The mean ± SD size of the pieces was of 28.4 ± 1.2 mm, and the time of ESD was 41.7 ± 2.4 min. The time of ESD in the first 15 procedures was 43.0 ± 3.0 min whereas in the next 15 procedures, the time was 40.3 ± 3.9 min, P = 0.588. The speed in the first 15 ESDs was 1.25 ± 0.11 cm(2)/min vs 2.12 ± 0.36 cm(2)/min in the remaining 15, P = 0.028. There were no complications. In patients, 5 lesions were resected en bloc. The size of the pieces was 25.2 ± 5.1 mm and the time was 85.0 ± 25.6 min. Endoscopic and histological controls did not show evidence of residual neoplastic tissue. A sequential ESD training program of a unique endoscopist, based on the practice in porcine models, contributed to learning ESD for its subsequent application in humans, yielding good results in efficacy and safety.

Efficient design of feedforward network for pattern classification

A feedforward neural network is a computing device whose processing units (the nodes) are distributed in adjacent layers connected through unidirectional links (the weights).Feedforward networks are widely used for pattern recognition. Here two feedforward networks are taken into consideration, Multi Layer Perceptron and Radial Basis Network. while designing these networks problem involves in finding the architecture which is efficient in terms of training time. In this paper different data samples will be presented to RBF and Multi Layer network and the best network selection will be done on the basis of minimum time taken by the network for training. Keywords- Feed forward network, Multi Layer Perceptron Neural Networks, Radial Basis Network, Spread. I. INTRODUCTION Pattern recognition is the study of how machines can observe the environment, learn to explore patterns of interest from their background, and make reasonable decisions about the classes of the patterns. the main properties of neural networks are that they have the ability to learn nonlinear input-output relationships, use sequential training procedures, and adapt themselves to the data. the most commonly used neural network family for pattern classification tasks is the feed-forward network, which includes Multi Layer Perceptron(MLP) and Radial-Basis Function (RBF) networks(9). Neural networks can be viewed as massively parallel computing systems consisting of an extremely large number of simple processors with many interconnections. Neural network models attempt to use some organizational principles (such as learning, generalization, adaptivity, fault tolerance and distributed representation. computation) in a network of weighted directed graphs in which the nodes are artificial neurons and directed edges (with weights) are connections between neuron outputs and neuron inputs. The main characteristics of neural networks are that they have the ability to learn complex nonlinear input-output relationships, use sequential training procedures, and adapt themselves to the data.

Impact of Bias-Correction Type and Conditional Training on Bayesian Model Averaging over the Northeast United States

Abstract The performance of a multimodel ensemble over the northeast United States is evaluated before and after applying bias correction and Bayesian model averaging (BMA). The 13-member Stony Brook University (SBU) ensemble at 0000 UTC is combined with the 21-member National Centers for Environmental Prediction (NCEP) Short-Range Ensemble Forecast (SREF) system at 2100 UTC. The ensemble is verified using 2-m temperature and 10-m wind speed for the 2007–09 warm seasons, and for subsets of days with high ozone and high fire threat. The impacts of training period, bias-correction method, and BMA are explored for these potentially hazardous weather events using the most recent consecutive (sequential training) and most recent similar days (conditional training). BMA sensitivity to the selection of ensemble members is explored. A running mean difference between forecasts and observations using the last 14 days is better at removing temperature bias than is a cumulative distribution function (CDF) or linear regression approach. Wind speed bias is better removed by adjusting the modeled CDF to the observation. High fire threat and ozone days exhibit a larger cool bias and a greater negative wind speed bias than the warm-season average. Conditional bias correction is generally better at removing temperature and wind speed biases than sequential training. Greater probabilistic skill is found for temperature using both conditional bias correction and BMA compared to sequential bias correction with or without BMA. Conditional and sequential BMA results are similar for 10-m wind speed, although BMA typically improves probabilistic skill regardless of training.

A Cervical Cerclage Task Trainer for Maternal-Fetal Medicine Fellows and Obstetrics/Gynecology Residents

Performance of cervical cerclage is a difficult procedure for learners with limited operative experience. We present a task trainer for cervical cerclage, which can facilitate learner training in a preclinical environment. A simulated vagina is first fashioned out of polyvinyl chloride pipe material to allow an appropriately shaped pelvic structure. To simulate collapsible vaginal sidewalls, a foam covered with latex sheets is used to line the inside of the task trainer. A "cervix" is made by cutting cylinders approximately 2 in in diameter and 4 in in length in frozen cow muscle. After thawing, the cervical models are fastened to a semirigid foam disk with suture, which is used to secure the cervix inside the polyvinyl chloride pipe at the top of the simulated vagina. The full cervical cerclage task trainer is then secured in a holder to stabilize it in place. Placement of a cervical cerclage can then be practiced with standard vaginal cerclage instruments. Cervical cerclage can be practiced in a nonclinical environment using this model. Several aspects of the procedure can be modified including the amount of vaginal tissue, length and caliber of the vagina, and pubic arch angle. This model allows beginning learners sequential training of increasing difficulty. Faculty and trainees alike found the trainer to be useful and felt that it helped in the acquisition of the surgical skills needed for cerclage placement in a clinical setting. This low-cost task trainer can provide Maternal-fetal Medicine fellows and obstetric/gynecology residents the opportunity to practice cervical cerclage placement in a nonclinical environment. It may help physicians to obtain or maintain proficiency despite the relative low frequency of the procedure in modern obstetric practice.

Regularized nonnegative matrix factorization using Gaussian mixture priors for supervised single channel source separation

We introduce a new regularized nonnegative matrix factorization (NMF) method for supervised single-channel source separation (SCSS). We propose a new multi-objective cost function which includes the conventional divergence term for the NMF together with a prior likelihood term. The first term measures the divergence between the observed data and the multiplication of basis and gains matrices. The novel second term encourages the log-normalized gain vectors of the NMF solution to increase their likelihood under a prior Gaussian mixture model (GMM) which is used to encourage the gains to follow certain patterns. In this model, the parameters to be estimated are the basis vectors, the gain vectors and the parameters of the GMM prior. We introduce two different ways to train the model parameters, sequential training and joint training. In sequential training, after finding the basis and gains matrices, the gains matrix is then used to train the prior GMM in a separate step. In joint training, within each NMF iteration the basis matrix, the gains matrix and the prior GMM parameters are updated jointly using the proposed regularized NMF. The normalization of the gains makes the prior models energy independent, which is an advantage as compared to earlier proposals. In addition, GMM is a much richer prior than the previously considered alternatives such as conjugate priors which may not represent the distribution of the gains in the best possible way. In the separation stage after observing the mixed signal, we use the proposed regularized cost function with a combined basis and the GMM priors for all sources that were learned from training data for each source. Only the gain vectors are estimated from the mixed data by minimizing the joint cost function. We introduce novel update rules that solve the optimization problem efficiently for the new regularized NMF problem. This optimization is challenging due to using energy normalization and GMM for prior modeling, which makes the problem highly nonlinear and non-convex. The experimental results show that the introduced methods improve the performance of single channel source separation for speech separation and speech–music separation with different NMF divergence functions. The experimental results also show that, using the GMM prior gives better separation results than using the conjugate prior.

A meta-cognitive sequential learning algorithm for neuro-fuzzy inference system

In this paper, we present a meta-cognitive sequential learning algorithm for a neuro-fuzzy inference system, referred to as, ‘Meta-Cognitive Neuro-Fuzzy Inference System’ (McFIS). McFIS has two components, viz., a cognitive component and a meta-cognitive component. The cognitive component employed is a Takagi–Sugeno–Kang type-0 neuro-fuzzy inference system. A self-regulatory learning mechanism that controls the learning process of the cognitive component, by deciding what-to-learn, when-to-learn and how-to-learn from sequential training data, forms the meta-cognitive component. McFIS realizes the above decision by employing sample deletion, sample reserve and sample learning strategy, respectively. The meta-cognitive component use the instantaneous error of the sample and spherical potential of the rule antecedents to select the best training strategy for the current sample. Also, in sample learning strategy, when a new rule is added the rule consequent is assigned such that the localization property of Gaussian rule is fully exploited. The performance of McFIS is evaluated on four regression and eight classification problems. The performance comparison shows the superior generalization performance of McFIS compared to existing algorithms.