Previous studies have confirmed that both honeybee and Drosophila are capable of learning and memory. This study aimed to investigate whether the house fly （Aldrichina grahami）, with strong instincts to adapt their ...Previous studies have confirmed that both honeybee and Drosophila are capable of learning and memory. This study aimed to investigate whether the house fly （Aldrichina grahami）, with strong instincts to adapt their living environment, have the learning ability to associate odor stimulus to avoid electric shock in free flying state using a device developed by the authors. The result showed the learning ability ofA. grahami at the electric shock voltages of 5 V, 25 V and 45 V AC. When 60 V was used, the flies were frequently injured. Our results indicate that A. grahami is a good model to study the neural mechanism of learning and memory. The paradigm in this study has some advantages that can be used in future studies of free insects.展开更多
. This paper proposes a novel remote sensing signal de-noising algorithm based on neural networks and tensor analysis. The defects exist in a constant deviation between the wavelet coeffi cients and that the wavelet c.... This paper proposes a novel remote sensing signal de-noising algorithm based on neural networks and tensor analysis. The defects exist in a constant deviation between the wavelet coeffi cients and that the wavelet coefficients of the noisy signal to estimate the discontinuity of hard threshold function and soft threshold function, limiting its further application in order to overcome this shortcoming, this paper proposes a new threshold function, compared with the original threshold function, a new threshold function is simple and easy to calculate, not only with the soft threshold function is continuous. To deal with this drawback, we integrate the NN to enhance the model. Neural network belongs to the basic unsupervised learning of neural networks, the principle of competition based on the mechanism of learning and biological and the memory capacity can be increased as the number of learning patterns increases, not only offi ine learning can also be carried out on-line ＂learning while learning＂ type. The integrated algorithm can host better performance.展开更多
文摘Previous studies have confirmed that both honeybee and Drosophila are capable of learning and memory. This study aimed to investigate whether the house fly （Aldrichina grahami）, with strong instincts to adapt their living environment, have the learning ability to associate odor stimulus to avoid electric shock in free flying state using a device developed by the authors. The result showed the learning ability ofA. grahami at the electric shock voltages of 5 V, 25 V and 45 V AC. When 60 V was used, the flies were frequently injured. Our results indicate that A. grahami is a good model to study the neural mechanism of learning and memory. The paradigm in this study has some advantages that can be used in future studies of free insects.
文摘. This paper proposes a novel remote sensing signal de-noising algorithm based on neural networks and tensor analysis. The defects exist in a constant deviation between the wavelet coeffi cients and that the wavelet coefficients of the noisy signal to estimate the discontinuity of hard threshold function and soft threshold function, limiting its further application in order to overcome this shortcoming, this paper proposes a new threshold function, compared with the original threshold function, a new threshold function is simple and easy to calculate, not only with the soft threshold function is continuous. To deal with this drawback, we integrate the NN to enhance the model. Neural network belongs to the basic unsupervised learning of neural networks, the principle of competition based on the mechanism of learning and biological and the memory capacity can be increased as the number of learning patterns increases, not only offi ine learning can also be carried out on-line ＂learning while learning＂ type. The integrated algorithm can host better performance.
