Variable three-term conjugate gradient method for training artificial neural networks.

Artificial neural networks (ANNs) have been widely adopted as general computational tools both in computer science as well as many other engineering fields. Stochastic gradient descent (SGD) and adaptive methods such as Adam are popular as robust optimization algorithms used to train the ANNs. However, the effectiveness of these algorithms is limited because they calculate a search direction based on a first-order gradient. Although higher-order gradient methods such as Newton's method have been proposed, they require the Hessian matrix to be semi-definite, and its inversion incurs a high computational cost. Therefore, in this paper, we propose a variable three-term conjugate gradient (VTTCG) method that approximates the Hessian matrix to enhance search direction and uses a variable step size to achieve improved convergence stability. To evaluate the performance of the VTTCG method, we train different ANNs on benchmark image classification and generation datasets. We also conduct a similar experiment in which a grasp generation and selection convolutional neural network (GGS-CNN) is trained to perform intelligent robotic grasping. After considering a simulated environment, we also test the GGS-CNN with a physical grasping robot. The experimental results show that the performance of the VTTCG method is superior to that of four conventional methods, including SGD, Adam, AMSGrad, and AdaBelief.

Investigations on the dissolved organic matter leached from oil-contaminated soils by using pyrolysis remediation method.

Pyrolysis, as a convenient and fast technology, has been proved to be promising in the remediation of oil-contaminated soil. However, little is known about the dissolved organic matter (DOM) associated with pyrolyzed oil-contaminated soil and its environmental impact. Herein, optical spectroscopic techniques (i.e., absorbance and fluorescence) were adopted to reveal the relationship between the pyrolysis temperature and the characteristics of the DOM and the associated phytotoxicity. Results show that one of the main factors determining the properties and phytotoxicity of DOM leached from the pyrolyzed soil is the critical temperature (approximately 325 °C) during pyrolysis. When the temperature was lower than 325 °C, more types and quantities of DOM, mainly fulvic acid-like substances, were desorbed from the soil with the temperature, which have little effect on wheat growth. However, when the temperature was in the range of 325-550 °C, the type and quantity of DOM increased first and then decreased as the temperature increased, during which the organic matter in the soil decomposed. The wheat growth was first inhibited and then promoted. Finally, the correlation between the spectral indices of DOM with the phytotoxicity suggested that fluorescent components identified by parallel factor analysis were positively correlated with phytotoxicity. This study indicates the pyrolytic remediation of oil-contaminated soil should avoid some critical temperature ranges.