Impact of Asymmetric Weight Update on Neural Network Training With Tiki-Taka Algorithm

Recent progress in novel non-volatile memory-based synaptic device technologies and their feasibility for matrix-vector multiplication (MVM) has ignited active research on implementing analog neural network training accelerators with resistive crosspoint arrays. While significant performance boost as well as area- and power-efficiency is theoretically predicted, the realization of such analog accelerators is largely limited by non-ideal switching characteristics of crosspoint elements. One of the most performance-limiting non-idealities is the conductance update asymmetry which is known to distort the actual weight change values away from the calculation by error back-propagation and, therefore, significantly deteriorates the neural network training performance. To address this issue by an algorithmic remedy, Tiki-Taka algorithm was proposed and shown to be effective for neural network training with asymmetric devices. However, a systematic analysis to reveal the required asymmetry specification to guarantee the neural network performance has been unexplored. Here, we quantitatively analyze the impact of update asymmetry on the neural network training performance when trained with Tiki-Taka algorithm by exploring the space of asymmetry and hyper-parameters and measuring the classification accuracy. We discover that the update asymmetry level of the auxiliary array affects the way the optimizer takes the importance of previous gradients, whereas that of main array affects the frequency of accepting those gradients. We propose a novel calibration method to find the optimal operating point in terms of device and network parameters. By searching over the hyper-parameter space of Tiki-Taka algorithm using interpolation and Gaussian filtering, we find the optimal hyper-parameters efficiently and reveal the optimal range of asymmetry, namely the asymmetry specification. Finally, we show that the analysis and calibration method be applicable to spiking neural networks.

Employee Performance Analysis Based on Human Resources Management Practices in Private Hospital in Surakarta

<p>Human resource management (HRM) is the foundation to face business competition. Organizational strategies related to HRM create satisfaction among employees so that employees perform well and can create a competitive advantage for their organization. The need for human resources in the hospital industry has developed along with the availability of hospitals reaching 2813 hospitals in Indonesia. This of course requires superior human resources performance. HRM practices are expected to drive human resources performance and achieve a competitive advantage in the hospital industry. This study will examine the effect of HRM practices, namely training, performance appraisal, and compensation systems on employee performance. Furthermore, this study will also examine the role of job satisfaction to mediate HRM practices on employee performance. The study was conducted in a private hospital in the Surakarta area with a sample of 120 respondents. Convenience sampling technique as a sampling method used to collect samples using a questionnaire. SmartPLS is used to analyze research data. The results showed that there was a significant positive effect on HRM Practices (training, performance appraisal, compensation system) on employee performance. However, job satisfaction only mediates the effect of performance appraisal and compensation system on employee performance, while job satisfaction does not play a role in mediating the effect of training on employee performance..</p>

Evaluating the Effects of Size and Precision of Training Data on ANN Training Performance for the Prediction of Chaotic Time Series Patterns

In this research, artificial neural networks (ANN) with various architectures are trained to generate the chaotic time series patterns of the Lorenz attractor. The ANN training performance is evaluated based on the size and precision of the training data. The nonlinear Auto-Regressive (NAR) model is trained in open loop mode first. The trained model is then used with closed loop feedback to predict the chaotic time series outputs. The research goal is to use the designed NAR ANN model for the simulation and analysis of Electroencephalogram (EEG) signals in order to study brain activities. A simple ANN topology with a single hidden layer of 3 to 16 neurons and 1 to 4 input delays is used. The training performance is measured by averaged mean square error. It is found that the training performance cannot be improved by solely increasing the training data size. However, the training performance can be improved by increasing the precision of the training data. This provides useful knowledge towards reducing the number of EEG data samples and corresponding acquisition time for prediction.

The Influence of Training, Performance Appraisal, and Organizational Support on Employee Service Behavior With Compensation as a Mediator

The training, performance appraisal, organizational support, and compensation are playing a very important role in enhancing employee service behavior in the police department. The main objective of this study is to examine the impact of training, performance appraisal, and organizational support on employee service behavior, with the mediating role of compensation. Data was collected from the employees of Punjab police through the questionnaire. The partial least square structural equation modeling (PLS-SEM) has been employed for data analysis. The results of the study found a positive and significant relationship between training, performance appraisal, organizational support, and employee service behavior. Furthermore, compensation plays the mediating role between training, performance appraisal, organizational support, and employee behavior in Punjab police, Pakistan. As per the findings of the study, employee behavior can be improved with the help of training, performance appraisal, organizational support, and compensation.

Dose Response of Acute ATP Supplementation on Strength Training Performance

Background: Chronic oral ATP supplementation benefits cardiovascular health, muscular performance, body composition, and recovery while attenuating muscle breakdown and fatigue. A single 400 mg dose of oral ATP supplementation improved lower body resistance training performance and energy expenditure in recreational resistance trained males, however, the minimal effective dose is currently unknown.Materials and Methods: Twenty recreationally trained men (age 28.6 ± 1.0 years, body mass 81.2 ± 2.0 kg, height 175.2 ± 1.4 cm, 1RM 141.5 ± 5.0 kg) consumed a single dose of either 400 mg, 200 mg, or 100 mg ATP (PEAK ATP®, TSI USA LLC, Missoula, MT, USA) or a placebo in a randomized, placebo-controlled crossover design, separated by a one week wash out between treatments. After warm-up, participants performed 4 sets of half-squats using free-weights until movement failure separated by 2 mins of rest between sets.Results: In comparison to placebo, 400 mg ATP significantly increased the number of set 1 repetitions (+13%, p = 0.04), and numerically increased total repetitions (+7%, p = 0.19) and total weight lifted (+6%, p = 0.22). 200 mg ATP numerically increased set 1 repetitions (+4% p = 0.47), while 100 mg ATP showed no improvements over placebo. 100 mg ATP (−4%, p &lt; 0.05) and 400 mg ATP (−4%, p = 0.11) decreased the perceived rate of exertion compared to placebo.Conclusions: In this study, the effective minimal dose of acute oral ATP supplementation during resistance exercise to increase performance was determined to be 400 mg, while as little as 100 mg showed improvements in perceived exertion.

An Improved Approach to Enhance Training Performance of ANN and the Prediction of PV Power for Any Time-Span without the Presence of Real-Time Weather Data

In this work, an improved approach to enhance the training performance of an Artificial Neural Network (ANN) for prediction of the output of renewable energy systems is proposed. Using the proposed approach, a significant reduction of the Mean Squared Error (MSE) in training performance is achieved, specifically from 4.45 × 10−7 to 3.19 × 10−10. Moreover, a simplified application of the already trained ANN is introduced through which photovoltaic (PV) output can be predicted without the availability of real-time current weather data. Moreover, unlike the existing prediction models, which ask the user to apply multiple inputs in order to forecast power, the proposed model requires only the set of dates specifying forecasting period as the input for prediction purposes. Moreover, in the presence of the historical weather data this model is able to predict PV power for different time spans rather than only for a fixed period. The prediction accuracy of the proposed model has been validated by comparing the predicted power values with the actual ones under different weather conditions. To calculate actual power, the data were obtained from the National Renewable Energy Laboratory (NREL), USA and from the Universiti Teknologi Malaysia (UTM), Malaysia. It is envisaged that the proposed model can be easily handled by a non-technical user to assess the feasibility of the photovoltaic solar energy system before its installation.

Impact of Emotional-Volitional Sphere of Future Specialists of Risky Professions on Professional Training Performance

Objective: The professional activities of specialists of risky professions are carried out in extreme conditions, characterized by a significant number of stressful factors, which, if the professional qualities of specialists are not sufficiently developed, leads to a decrease in the efficiency of performance of official duties, professional burnout, and psychosomatic disorders. Background: The success of the performance of service tasks is primarily influenced by the developmental level of the emotional-volitional sphere, in particular self-regulation, emotional-volitional stability, ensuring the ability to make autonomous decisions, form and determine tasks following the requirements of complex, changing situations, and therefore, achieve their goals. The article aims to analyze the emotional and volitional state of future specialists of risky professions during training. Method: The leading research method was observation. The emotional-volitional sphere is an integral part of the system of regulating activity as a professional. During the research, the main areas of the formation of the emotional-volitional sphere were identified. With the help of psychological analysis, the important properties of future specialists in the development of the emotional-volitional sphere have been determined. Results: The analysis results show that future specialists have an average and low level of stylistic possibilities for self-regulation (modeling, programming, autonomy). The reasons for the low progress of future specialists have been determined. Conclusion: The practical significance of the research lies in the development of recommendations for training and the formation of the emotional-volitional sphere during educational activities.