Hard Training Research Articles

A low cost neuromorphic learning engine based on a high performance supervised SNN learning algorithm

Spiking neural networks (SNNs) are more energy- and resource-efficient than artificial neural networks (ANNs). However, supervised SNN learning is a challenging task due to non-differentiability of spikes and computation of complex terms. Moreover, the design of SNN learning engines is not an easy task due to limited hardware resources and tight energy constraints. In this article, a novel hardware-efficient SNN back-propagation scheme that offers fast convergence is proposed. The learning scheme does not require any complex operation such as error normalization and weight-threshold balancing, and can achieve an accuracy of around 97.5% on MNIST dataset using only 158,800 synapses. The multiplier-less inference engine trained using the proposed hard sigmoid SNN training (HaSiST) scheme can operate at a frequency of 135 MHz and consumes only 1.03 slice registers per synapse, 2.8 slice look-up tables, and can infer about 0.03times {varvec{10}}^{varvec{9}} features in a second, equivalent to 9.44 giga synaptic operations per second (GSOPS). The article also presents a high-speed, cost-efficient SNN training engine that consumes only 2.63 slice registers per synapse, 37.84 slice look-up tables per synapse, and can operate at a maximum computational frequency of around 50 MHz on a Virtex 6 FPGA.

ANALYSIS OF ADJECTIVES TSURAI, KURUSHII AND KUTSUU IN JAPANESE LANGUAGE RUIGIGO

Japanese is a foreign language that has many variations of the lexicon. One of them is the lexicon variation in adjectives. This article discusses the analysis of the adjectives tsurai, kutsuhii and kutsuu in Japanese ruigigo. This study focuses on sentences used by native Japanese speakers. Data collection was carried out through interviews and note-taking techniques. The author conducts interviews directly with Japanese people while recording sentences containing 'tsurai, kutsuhii and kutsuu'. The data were analyzed using a contextual meaning theory approach (Pateda, 2010) by describing them in Japanese sentences. From the data found, the adjectives tsurai, kutsuhii and kutsuu have different meanings depending on the context of their usage. Based on the results of the analysis, in general (1) tsurai expressed a bitter experience, something uncomfortable, lazy to do something, (2) kutsuu stated difficult family conditions, difficult life, doing hard training), and (3) kutsuu expressed severe pain physical and mental discomfort and torture. The results of the study show that each lexicon of 'tsurai, kutsuhii and kutsuu' has almost the same meaning, namely painful or difficult, but has a slightly different usage in other activities.

Upaya Mereduksi Aksi Klitih oleh Pelajar Melalui Pelatihan Pemberdayaan Psikis dan Perilaku di Yogyakarta

This community dedication aims to provide education regarding the prevention of klitih through psychological and behavioral empowerment which is implemented through the Sekolah Anti Klitih (SETITI). This activity is surely the result of collaboration with partner, in this case the partner involved is MAN 3 Sleman which is located on Jalan Magelang km. 4, Sinduadi, Mlati, Sleman, Special Region of Yogyakarta. The object determined by the partner leadership, in this case the guidance and counseling teacher of MAN 3 Sleman who is also a Scoutmaster, is the students of MAN 3 Sleman who are members of the Scouting Ambalan Council. This activity consists of training and mentoring programs. The series of activities in the training program are SETITI program socialization, focus group discussions, and soft and hard skills training. Meanwhile, the series of mentoring activities program are sharing sessions, dream posters, and SETITI festival as the culmination of the activities of this community dedication program. Increased knowledge of students related to klitih can be seen through the results of the pre-test and post-test, as well as from the behavior of students who are more disciplined after participating in a series of activities. It is hoped that the improvement shown by the target audience can be a step for the formation of klitih preventive agents.

Systematic Review of Smart Classroom for Hard Skills Training in Augmented and Virtual Reality Environments

Advances in augmented and virtual reality (AVR) technology have allowed for the development of AVR interactive learning environments (AVR-ILEs) with increasing fidelity. When paired with a suitably capable computer tutor agent, such environments can permit adaptive and self-directed learning of procedural skills in some cases. We undertook a comprehensive review of published smart classroom (SC) solutions to facilitate hard skills training in AVR-ILEs. We established the learning context and implemented support for hard skills training for each of the seven solutions that qualified for the final analysis. Typically, AVR-ILEs mimic realistic work conditions or equipment. The aim of this study is to solutions featuring a smart classroom for hard skills training in an AVR-ILEs have been described. In the study, it is also aimed to types of AVR-ILEs from a smart classroom (SC) have been implemented to support hard skills training. This systematic review is based on the reporting criteria of PRISMA 2020 and the scope of the study consists of 9 academic studies including the keyword "smart classroom" and the keywords "augmented reality" or "virtual reality" in academic studies scanned in Scopus, IEEE Xplore, Eric, ScienceDirect and ACM Digital Library. In this study, the distribution of the studies used in the Smart Classroom for Hard Skills Training in Augmented and Virtual Reality Environments (SCHST-AVRE) by solution, domain, country, concept description, country, reviewed papers, learning objective(s), learning tasks, and the context of the technology used is included. In order to contribute to the research on the AVR-ILEs, it has been tried to determine the opportunities and risks that teaching in the SCHST-AVRE can offer to the field of education. In addition, it is thought that this study will shed a light on future studies.

Effects of Rowing on Rheological Properties of Blood

The aim of this study was to analyze the selected hematological and rheological indices in female rowers during the competitive season. The study included 10 female rowers (aged 21.2 ± 2.6) and the control group consisted of 10 woman of corresponding age (non-athletes). The examination of athletes took place two times: at the beginning of the season during high endurance low intensity training period in January (baseline) and at the end of the competitive season in October (after). Blood samples taken from all woman were analyzed for hematological and rheological parameters. The training period of rowers during the 10 months resulted in decrease in red blood cell count and RBC deformability, in contrast to an improvement in some rheological functions such a decrease in fibrinogen concentration, plasma viscosity and aggregation index. The training program practice in rowing modulated some hematological and rheological indices. Some of them positively influenced the cardiovascular system and reduced potential risks connected with hard training and dehydration, but others may have followed from overtraining or not enough relaxation time between training units.

The Inculcative Power of Australian Cadet Corps Uniforms in the 1900s and 1910s

The Inculcative Power of Australian Cadet Corps Uniforms in the 1900s and 1910s

Gamifying model-based engineering: the PapyGame experience

Modeling is an essential and challenging activity in any engineering environment. It implies some hard-to-train skills such as abstraction and communication. Teachers, project leaders, and tool vendors have a hard time teaching or training their students, co-workers, or users. Gamification refers to the exploitation of gaming mechanisms for serious purposes, like promoting behavioral changes, soliciting participation and engagement in activities, etc. We investigate the introduction of gaming mechanisms in modeling tasks with the primary goal of supporting learning/training. The result has been the realization of a gamified modeling environment named PapyGame. In this article, we present the approach adopted for PapyGame implementation, the details on the gamification elements involved, and the derived conceptual architecture required for applying gamification in any modeling environment. Moreover, to demonstrate the benefits of using PapyGame for learning/training modeling, a set of user experience evaluations have been conducted. Correspondingly, we report the obtained results together with a set of future challenges we consider as critical to make gamified modeling a more effective education/training approach.

BALAI LATIHAN KERJA (BLK) DALAM MENINGKATKAN CIVIC COMPETENCE TENAGA KERJA DI ERA REVOLUSI 4.0

This article aims to explain the increase of civic competence for the workforce in vocational training centers (BLK) in the era of the industrial revolution 4.0. This research uses a qualitative descriptive method. This research was conducted at BLK level 1 UPTP Aceh. Sources of data obtained by observation, interviews, and document analysis. Data analysis uses mile and hubberman by data reduction, triangulation and display. The results of the study show that BLK have carried out civic competence improvements through soft skill and hard skill training programs and adapted to the 4.0 industrial revolution era. This can be seen through the determination of programs that use TNA (training need analysis), its means looking at programs by developments needs of the labor market before opening the type of training programs and based on DUDI (business and industry).

Multi-Label Hashing for Dependency Relations among Multiple Objectives.

Learning hash functions have been widely applied for large-scale image retrieval. Existing methods usually use CNNs to process an entire image at once, which is efficient for single-label images but not for multi-label images. First, these methods cannot fully exploit independent features of different objects in one image, resulting in some small object features with important information being ignored. Second, the methods cannot capture different semantic information from dependency relations among objects. Third, the existing methods ignore the impacts of imbalance between hard and easy training pairs, resulting in suboptimal hash codes. To address these issues, we propose a novel deep hashing method, termed multi-label hashing for dependency relations among multiple objectives (DRMH). We first utilize an object detection network to extract object feature representations to avoid ignoring small object features and then fuse object visual features with position features and further capture dependency relations among objects using a self-attention mechanism. In addition, we design a weighted pairwise hash loss to solve the imbalance problem between hard and easy training pairs. Extensive experiments are conducted on multi-label datasets and zero-shot datasets, and the proposed DRMH outperforms many state-of-the-art hashing methods with respect to different evaluation metrics.

General Cyclical Training of Neural Networks

This position paper describes the principle of “General Cyclical Training” in machine learning, where training starts and ends with “easy training” and the “hard training” happens during the middle epochs. We propose several manifestations for training neural networks, including algorithmic examples (via hyper-parameters and loss functions), data-based examples, and model-based examples. Specifically, we introduce several novel techniques: cyclical weight decay, cyclical batch size, cyclical focal loss, cyclical softmax temperature, cyclical data augmentation, cyclical gradient clipping, and cyclical semi-supervised learning. In addition, we demonstrate that cyclical weight decay, cyclical softmax temperature, and cyclical gradient clipping (as three examples of this principle) are beneficial in the test accuracy performance of a trained model. Furthermore, we discuss model-based examples (such as pretraining and knowledge distillation) from the perspective of general cyclical training and recommend some changes to the typical training methodology. In summary, this paper defines the general cyclical training concept and discusses several specific ways in which this concept can be applied to training neural networks. In the spirit of reproducibility, the code used in our experiments is available at https://github.com/lnsmith54/CFL.

Intra-Class Adaptive Augmentation With Neighbor Correction for Deep Metric Learning

Deep metric learning aims to learn an embedding space, where semantically similar samples are close together and dissimilar ones are repelled against. To explore more hard and informative training signals for augmentation and generalization, recent methods focus on generating synthetic samples to boost metric learning losses. However, these methods just use the deterministic and class-independent generations (e.g., simple linear interpolation), which only can cover the limited part of distribution spaces around original samples. They have overlooked the wide characteristic changes of different classes and can not model abundant intra-class variations for generations. Therefore, generated samples not only lack rich semantics within the certain class, but also might be noisy signals to disturb training. In this paper, we propose a novel intra-class adaptive augmentation (IAA) framework for deep metric learning. We reasonably estimate intra-class variations for every class and generate adaptive synthetic samples to support hard samples mining and boost metric learning losses. Further, for most datasets that have a few samples within the class, we propose the neighbor correction to revise the inaccurate estimations, according to our correlation discovery where similar classes generally have similar variation distributions. Extensive experiments on five benchmarks show our method significantly improves and outperforms the state-of-the-art methods on retrieval performances by 3%-6%. Our code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/darkpromise98/IAA</uri> .

OAF-Net: An Occlusion-Aware Anchor-Free Network for Pedestrian Detection in a Crowd

Although pedestrian detection has achieved promising performance with the development of deep learning techniques, it remains a great challenge to detect heavily occluded pedestrians in crowd scenes. Therefore, to make the anchor-free network pay more attention to learning the hard examples of occluded pedestrians, we propose a simple but effective Occlusion-aware Anchor-Free Network (namely OAF-Net) for pedestrian detection in crowd scenes. Specifically, we first design a novel occlusion-aware detection head, which includes three separate center prediction branches combining with the scale and offset prediction branches. In the detection head of OAF-Net, occluded pedestrian instances are assigned to the most suitable center prediction branch according to the occlusion level of human body. To optimize the center prediction, we accordingly propose a novel weighted Focal Loss where pedestrian instances are assigned with different weights according to their visibility ratios, so that the occluded pedestrians are up-weighted during the training process. Our OAF-Net is able to model different occlusion levels of pedestrian instances effectively, and can be optimized towards catching a high-level understanding of the hard training samples of occluded pedestrians. Experiments on the challenging CityPersons, Caltech, and CrowdHuman benchmarks sufficiently validate the efficacy of our OAF-Net for pedestrian detection in crowd scenes.

Tohjm-Trained Multiscale Spatial Temporal Graph Convolutional Neural Network for Semi-Supervised Skeletal Action Recognition

In recent years, spatial-temporal graph convolutional networks have played an increasingly important role in skeleton-based human action recognition. However, there are still three major limitations to most ST-GCN-based approaches: (1) They only use a single joint scale to extract action features, or process joint and skeletal information separately. As a result, action features cannot be extracted dynamically through the mutual directivity between the scales. (2) These models treat the contributions of all joints equally in training, which neglects the problem that some joints with difficult loss-reduction are critical joints in network training. (3) These networks rely heavily on a large amount of labeled data, which remains costly. To address these problems, we propose a Tohjm-trained multiscale spatial-temporal graph convolutional neural network for semi-supervised action recognition, which contains three parts: encoder, decoder and classifier. The encoder’s core is a correlated joint–bone–body-part fusion spatial-temporal graph convolutional network that allows the network to learn more stable action features between coarse and fine scales. The decoder uses a self-supervised training method with a motion prediction head, which enables the network to extract action features using unlabeled data so that the network can achieve semi-supervised learning. In addition, the network is also capable of fully supervised learning with the encoder, decoder and classifier. Our proposed time-level online hard joint mining strategy is also used in the decoder training process, which allows the network to focus on hard training joints and improve the overall network performance. Experimental results on the NTU-RGB + D dataset and the Kinetics-skeleton dataset show that the improved model achieves good performance for action recognition based on semi-supervised training, and is also applicable to the fully supervised approach.

Differentiating Stroke and Movement Accelerometer Profiles to Improve Prescription of Tennis Training Drills.

Perri, T, Reid, M, Murphy, A, Howle, K, and Duffield, R. Differentiating stroke and movement accelerometer profiles to improve prescription of tennis training drills. J Strength Cond Res 37(3): 646-651, 2023-This study compared the movement- and stroke-related accelerometer profiles and stroke counts between common on-court tennis training drills. Ten, junior-elite, male tennis players wore a cervical-mounted global positioning systems, with in-built accelerometer, gyroscope, and magnetometer during hard court training sessions ( n = 189). Individual training drills were classified into 8 categories based on previous research descriptions. Manufacturer software calculated total player load (tPL), while a prototype algorithm detected forehand (FH), backhands (BH), and serves and then calculated a stroke player load (sPL) from individual strokes. Movement player load (mPL) was calculated as the difference between tPL and sPL. Drill categories were compared for relative ( . min -1 ) tPL, sPL, mPL, and stroke counts via a 1-way analysis of variance with effect sizes (Cohen's d ) and 95% confidence intervals. Highest tPL . min -1 existed in accuracy and recovery or defensive drills ( p < 0.05), with lowest tPL·min -1 values observed in match-play simulation ( p < 0.05). For sPL·min -1 , accuracy drills elicited greater values compared with all other drill types ( p < 0.05), partly via greater FH-sPL·min -1 ( p < 0.05), with lowest sPL·min -1 existing for match-play ( p < 0.05). Accuracy, open, and recovery or defensive drills result in greater BH-sPL·min -1 and BH . min -1 ( p < 0.05). Serve-sPL·min -1 is highest in technical and match-play drills ( p < 0.05). Higher mPL·min -1 existed in accuracy, recovery or defensive, 2v1 net, open, and 2v1 baseline ( p < 0.05). Furthermore, mPL·min -1 in points drills was greater than technical and match-play simulation drills ( p < 0.05). Higher hitting-based accelerometer loads (sPL·min -1 ) exist in accuracy drills, whereas technical and match-play drills show the lowest movement demands (mPL·min -1 ). These findings can aid individual drill prescription for targeting movement or hitting load.

Interval-Valued Reduced RNN for Fault Detection and Diagnosis for Wind Energy Conversion Systems

Recurrent neural network (RNN) is one of the most used deep learning techniques in fault detection and diagnosis (FDD) of industrial systems. However, its implementation suffers from some limitations presented in the hard training step and the high time complexity. Besides, most used RNN-based FDD techniques do not deal with system uncertainties. Therefore, this paper proposes enhanced RNN techniques that detect and classify faults in wind energy conversion (WEC) systems. First, we develop a reduced RNN in order to simplify the model in terms of training and complexity time as well. Reduced RNN is based on Hierarchical K-means clustering to treat the correlations between samples and extract a reduced number of observations from the training data matrix. Second, two reduced RNN-based interval-valued-data techniques are proposed to distinguish between the different WEC system operating modes. The proposed techniques for interval-valued data are able to improve both fault diagnosis robustness and susceptibility while maintaining a satisfactory and stable performance over long periods of process operation. The presented results confirm the high feasibility and effectiveness of the proposed FDD techniques (an accuracy greater than 98% for all the proposed methods).

MOTIVATIONAL FACTORS ASSOCIATED WITH SPORTS PARTICIPATION AMONG ATHLETES

Background: Successful athletes are motivated for future-oriented for more achievements. Apprehensive athletes are not as successful as highly motivated athletes. The sport participation is influenced by many factors those contribute in success of athletes. Objective: To explore the motivational factors for sports participation among Pakistani athletes. Methodology: The cross-sectional survey was conducted on both male and female athletes (n=400) between the age of 14 to 40 years, played at the national and international levels at least from one year. The study was conducted at the Pakistan sports board (PSB) from July-December 2018. The non-probability convenience sampling was used in this study. The sports participation scale-6 was used to explore the motivational factors. The data was presented as n(%). Results: the mean age of the participants were 24±5.45 years The results showed that the mean scores of six categories of Sports Motivational Scale (SMS) was 134.69±13.431, which showed that majority of athletes were highly motivated. The result of the study also showed that all subdomain of sports motivation scale positively correlated to total score (p&lt;0.001), while a motivation was negatively associated r=-0.701, p&lt;0.001) with total score of sports motivation scale. Although most of the athletes were normal in term of depression, anxiety, and stress. But depression is negatively correlated (r=-0.09, p=0.04) with sports motivation. Conclusion: This study concluded that most athletes are motivated because hard training improves their performance and other motivational factors are also associated like pleasure, internal satisfaction, and excitement. The least motivational factor was achieving social or material benefits of being an athlete. Keywords: Athlete, motivation, participation, stress.

Improving Home-Based Physical Therapy Treatments: Examining the Doctor-Patient Relationship

While recent advancements have been made regarding physical therapy techniques and effectiveness, patient adherence to prescribed home exercise programs (HEPs)--a critical piece of successful physical therapy outcomes--remains low. There is growing evidence that establishing a good relationship and line of communication with patients is just as important as hard skill medical training when trying to improve patient adherence to treatment regimens. As more focus is given to the “human” elements of physical therapy, a vast array of training methods and programs have emerged as educators attempt to prepare their students for a career heavily contingent on creating healthy relationships between patients and their therapist. Unlike previous research meant to evaluate the efficacy of such programs, this study aims through semi-structured interviews to (a) examine the availability of such training in a particular geographic area, and (b) serve as a baseline for future research to gain a comprehensive understanding of the strengths and weaknesses different regions have in terms of relationship and communication training. Findings indicate that in Southwest Missouri, adherence to HEPs matches previous data and that little explicit educational focus is given to improving the doctor-patient relationship. However, interviewees indicated that there are many implicit elements in their education that effectively improved their relationship skills, and that explicit focus is steadily increasing. This study is successful in taking the first step in evaluating what it will take to standardize relationship and communication education among physical therapists.

Hard Boundary-Based Neurofeedback Training Procedure: A Modified Fixed Thresholding Method for More Accurate Guidance of Subjects Within Target Areas During Neurofeedback Training.

In nearly all studies within the domain of neurofeedback, a threshold has been defined for each training feature in a way that subjects' status can be evaluated during training according to the given value. In this study, a hard boundary-based neurofeedback training (HBNFT) method based on the determination of decision boundary using support vector machine (SVM) classifier was proposed in which subjects' status were clarified considering a decision boundary and they could also be encouraged once entering a target area. In this method, a scoring index (SI) was similarly defined whose value was determined in accordance with subject performance during training. The results revealed that employing a classifier and determining a decision boundary instead of using a threshold could prove more successful in accurately guiding them towards a target area and also meet no needs to choose a basis for determining a threshold. Moreover, it was likely that the proposed method could be more efficient in controlling features and preventing extreme changes compared to those using variable thresholds.

THE MONITORING OF NON-FARMING AND NON-GRAIN PURPOSES IN ARABLE LAND OF ZHEJIANG, CHINA WITH DOMESTIC SATELLITE IMAGERY

Abstract. Arable land protection is essential for Zero hunger the Sustainable Development Goals of United Nations, and the arable land protection includes two aspects, non-farming and non-grain. We try to monitor the arable land protection in Zhejiang with domestic satellite imagery. Satellite remote technology has become an essential way to monitor the land cover change (for non-farming) and grain crops (for non-grain). However, current monitoring frequency and scale were unable to satisfy the needs for non-farming monitoring. The low-resolution image cannot satisfy the feature of land fragmentation of Zhejiang for non-grain monitoring. To address the above problem, this paper proposes a land cover change method to monitor non-farming purposes based on deeplabv3+ with monthly coverage 2 meters resolution images. By focusing on rebuilding training data set and improving training strategy with hard example training, the difficulty of the spurious change caused by the adjustment of farming structure is solved. At the same time, this paper builds three training processes (Initial training, Fine training, Retraining for promotion) based on Fully Convolutional Neural Network FCN-8S to monitor the main grain crops in Zhejiang. The phenological features are added into the process of training to further improve the accuracy. At present, land cover change method of this paper has been applied in Zhejiang province and the monitoring of grain crops has been carried out in some regions according to the specific requirements. The result shows that both the two methods exhibit good accuracy and generalization ability at the time and space scale.

Systematic Review of Intelligent Tutoring Systems for Hard Skills Training in Virtual Reality Environments

Advances in immersive virtual reality (I-VR) technology have allowed for the development of I-VR learning environments (I-VRLEs) with increasing fidelity. When coupled with a sufficiently advanced computer tutor agent, such environments can facilitate asynchronous and self-regulated approaches to learning procedural skills in industrial settings. In this study, we performed a systematic review of published solutions involving the use of an intelligent tutoring system (ITS) to support hard skills training in an I-VRLE. For the seven solutions that qualified for the final analysis, we identified the learning context, the implemented system, as well as the perceptual, cognitive, and guidance features of the utilized tutoring agent. Generally, the I-VRLEs emulated realistic work environments or equipment. The solutions featured either embodied or embedded tutor agents. The agents’ perception was primarily based on either learner actions or learner progress. The agents’ guidance actions varied among the solutions, ranging from simple procedural hints to event interjections. Several agents were capable of answering certain specific questions. The cognition of the majority of agents represented variations on branched programming. A central limitation of all the solutions was that none of the reports detailed empirical studies conducted to compare the effectiveness of the developed training and tutoring solutions.