scholarly journals A Novel Training Protocol for Performance Predictors of Evolutionary Neural Architecture Search Algorithms

2021 ◽  
pp. 1-1
Author(s):  
Yanan Sun ◽  
Xian Sun ◽  
Yuhan Fang ◽  
Gary G. Yen ◽  
Yuqiao Liu
Keyword(s):  
Search Algorithms ◽  
Neural Architecture ◽  
Performance Predictors ◽  
Training Protocol

scholarly journals TextNAS: A Neural Architecture Search Space Tailored for Text Representation

2020 ◽  
Vol 34 (05) ◽  
pp. 9242-9249
Author(s):  
Yujing Wang ◽  
Yaming Yang ◽  
Yiren Chen ◽  
Jing Bai ◽  
Ce Zhang ◽  
...  
Keyword(s):  
Text Classification ◽  
Network Architecture ◽  
State Of The Art ◽  
Search Space ◽  
Search Algorithms ◽  
Text Representation ◽  
Neural Architecture ◽  
Good Potential ◽  
Automatic Search ◽  
Public Datasets

Learning text representation is crucial for text classification and other language related tasks. There are a diverse set of text representation networks in the literature, and how to find the optimal one is a non-trivial problem. Recently, the emerging Neural Architecture Search (NAS) techniques have demonstrated good potential to solve the problem. Nevertheless, most of the existing works of NAS focus on the search algorithms and pay little attention to the search space. In this paper, we argue that the search space is also an important human prior to the success of NAS in different applications. Thus, we propose a novel search space tailored for text representation. Through automatic search, the discovered network architecture outperforms state-of-the-art models on various public datasets on text classification and natural language inference tasks. Furthermore, some of the design principles found in the automatic network agree well with human intuition.

scholarly journals Evolution of Activation Functions: An Empirical Investigation

2021 ◽  
Vol 1 (2) ◽  
pp. 1-36
Author(s):  
Andrew Nader ◽  
Danielle Azar
Keyword(s):  
Neural Network ◽  
Empirical Investigation ◽  
Expert Knowledge ◽  
Search Algorithms ◽  
Trial And Error ◽  
Activation Functions ◽  
Neural Architecture ◽  
Out Of The Loop ◽  
Good Set ◽  
Hidden Neurons

The hyper-parameters of a neural network are traditionally designed through a time-consuming process of trial and error that requires substantial expert knowledge. Neural Architecture Search algorithms aim to take the human out of the loop by automatically finding a good set of hyper-parameters for the problem at hand. These algorithms have mostly focused on hyper-parameters such as the architectural configurations of the hidden layers and the connectivity of the hidden neurons, but there has been relatively little work on automating the search for completely new activation functions, which are one of the most crucial hyperparameters to choose. There are some widely used activation functions nowadays that are simple and work well, but nonetheless, there has been some interest in finding better activation functions. The work in the literature has mostly focused on designing new activation functions by hand or choosing from a set of predefined functions while this work presents an evolutionary algorithm to automate the search for completely new activation functions. We compare these new evolved activation functions to other existing and commonly used activation functions. The results are favorable and are obtained from averaging the performance of the activation functions found over 30 runs, with experiments being conducted on 10 different datasets and architectures to ensure the statistical robustness of the study.

Making targeted dimensions transparent on relations with typical performance predictors

2007 ◽  
Author(s):  
Kimberly A. Smith-Jentsch ◽  
Helen M. Boudreau ◽  
Michael Flood
Keyword(s):  
Performance Predictors ◽  
Typical Performance

Knowledge training versus process training: The effects of training protocol on team coordination and performance

2007 ◽  
Author(s):  
Jamie C. Gorman ◽  
Nancy J. Cooke ◽  
Jennifer L. Winner ◽  
Jasmine L. Duran ◽  
Harry K. Pedersen ◽  
...  
Keyword(s):  
Team Coordination ◽  
And Performance ◽  
Training Protocol

A Neural Architecture for Visual Search

1992 ◽  
Author(s):  
William Ross ◽  
Ennio Mingolla
Keyword(s):  
Visual Search ◽  
Neural Architecture

Physical exercise protects dynamic balance and motor coordination of rats treated with vincristine

2020 ◽  
Vol 19 (5) ◽  
pp. 336
Author(s):  
Luiza Minato Sagrillo ◽  
Viviane Nogueira De Zorzi ◽  
Luiz Fernando Freire Royes ◽  
Michele Rechia Fighera ◽  
Beatriz Da Silva Rosa Bonadiman ◽  
...  
Keyword(s):  
Oxidative Stress ◽  
Locomotor Activity ◽  
Physical Exercise ◽  
Motor Coordination ◽  
Dynamic Balance ◽  
Exercise Group ◽  
Adult Rats ◽  
The Central Nervous System ◽  
Stress Damage ◽  
Training Protocol

Physical exercise has been shown to be an important modulator of the antioxidant system and neuroprotective in several diseases and treatments that affect the central nervous system. In this sense, the present study aimed to evaluate the effect of physical exercise in dynamic balance, motor coordination, exploratory locomotor activity and in the oxidative and immunological balance of rats treated with vincristine (VCR). For that, 40 adult rats were divided into two groups: exercise group (6 weeks of swimming, 1h/day, 5 days/week, with overload of 5% of body weight) and sedentary group. After training, rats were treated with 0.5 mg/kg of vincristine sulfate for two weeks or with the same dose of 0.9% NaCl. The behavioral tests were conducted 1 and 7 days after each dose of VCR. On day 15 we carried out the biochemical analyzes of the cerebellum. The physical exercise was able to protect against the loss of dynamic balance and motor coordination and, had effect per se in the exploratory locomotor activity, and neutralize oxidative stress, damage DNA and immune damage caused by VCR up to 15 days after the end of the training protocol. In conclusion, we observed that previous physical training protects of the damage motor induced by vincristine.Key-words: exercise, oxidative stress, neuroprotection, cerebellum.

The influence of execution speed on blood lactate concentration in strength training protocol in bench press exercise

2020 ◽  
Vol 19 (1) ◽  
pp. 32
Author(s):  
Gustavo Taques Marczynski ◽  
Luís Carlos Zattar Coelho ◽  
Leonardo Emmanuel De Medeiros Lima ◽  
Rodrigo Pereira Da Silva ◽  
Dilmar Pinto Guedes Jr ◽  
...  
Keyword(s):  
Strength Training ◽  
Blood Lactate ◽  
Bench Press ◽  
Lactate Concentration ◽  
Blood Lactate Concentration ◽  
The Third ◽  
Fast Speed ◽  
Work Volume ◽  
Execution Speed ◽  
Training Protocol

The aim of this study was to analyze the influence of two velocities of execution relative to blood lactate concentration in strength training exercise until the momentary concentric failure. Fifteen men (29.1 ± 5.9 years), trained, participated in the experiment. The volunteers performed three bench press sessions, with an interval of 48 hours between them. At the first session, individuals determined loads through the 10-12 RMs test. In the following two sessions, three series with 90 seconds of interval were performed, in the second session slow execution speed (cadence 3030) and later in the third session fast speed (cadence 1010). For statistical analysis, the Student-T test was used for an independent sample study and considered the value of probability (p) ≤ 0.05 statistically significant. By comparing the number of repetitions and time under tension of the two runs, all series compared to the first presented significant reductions (p < 0.05). The total work volume was higher with the fast speed (p < 0.05). The study revealed that rapid velocities (cadence 1010) present a higher concentration of blood lactate when compared to slow runs (cadence 3030). The blood lactate concentration, in maximum repetitions, is affected by the speed of execution.Keywords: resistance training, cadence, blood lactate.

Psychophysiological Responses of Pilots in Hypoxia Training at 7000 and 7500 m

2020 ◽  
Vol 91 (10) ◽  
pp. 785-789
Author(s):  
Dongqing Wen ◽  
Lei Tu ◽  
Guiyou Wang ◽  
Zhao Gu ◽  
Weiru Shi ◽  
...  
Keyword(s):  
Heart Rate ◽  
Psychomotor Performance ◽  
Physiological Responses ◽  
Arterial Oxygen Saturation ◽  
Arterial Oxygen ◽  
Awareness Training ◽  
Similar Frequency ◽  
Psychophysiological Responses ◽  
The Mean ◽  
Training Protocol

INTRODUCTION: We compared the physiological responses, psychomotor performances, and hypoxia symptoms between 7000 m and 7500 m (23,000 and 24,600 ft) exposure to develop a safer hypoxia training protocol.METHODS: In altitude chamber, 66 male pilots were exposed to 7000 and 7500 m. Heart rate and arterial oxygen saturation were continuously monitored. Psychomotor performance was assessed using the computational task. The hypoxic symptoms were investigated by a questionnaire.RESULTS: The mean duration time of hypoxia was 323.0 56.5 s at 7000 m and 218.2 63.3 s at 7500 m. The 6-min hypoxia training was completed by 57.6% of the pilots and 6.1% of the pilots at 7000 m and at 7500 m, respectively. There were no significant differences in pilots heart rates and psychomotor performance between the two exposures. The Spo2 response at 7500 m was slightly severer than that at 7000 m. During the 7000 m exposure, pilots experienced almost the same symptoms and similar frequency order as those during the 7500 m exposure.CONCLUSIONS: There were concordant symptoms, psychomotor performance, and very similar physiological responses between 7000 m and 7500 m during hypoxia training. The results indicated that 7000-m hypoxia awareness training might be an alternative to 7500-m hypoxia training with lower DCS risk and longer experience time.Wen D, Tu L, Wang G, Gu Z, Shi W, Liu X. Psychophysiological responses of pilots in hypoxia training at 7000 and 7500 m. Aerosp Med Hum Perform. 2020; 91(10):785789.

Sign in / Sign up

Export Citation Format

Share Document