Depeng Gao, Jiafeng Liu, Rui Wu, Dansong Cheng, Xiaopeng Fan and Xianglong Tang
With the advent of 3D cameras, getting depth information along with RGB images has been facilitated, which is helpful in various computer vision tasks. However, there are two challenges in using these RGB-D images to help recognize RGB images captured by conventional cameras: one is that the depth images are missing at the testing stage, the other is that the training and test data are drawn from different distributions as they are captured using different equipment. To jointly address the two challenges, we propose an asymmetrical transfer learning framework, wherein three classifiers are trained using the RGB and depth images in the source domain and RGB images in the target domain with a structural risk minimization criterion and regularization theory. A cross-modality co-regularizer is used to restrict the two-source classifier in a consistent manner to increase accuracy. Moreover, an L2,1 norm cross-domain co-regularizer is used to magnify significant visual features and inhibit insignificant ones in the weight vectors of the two RGB classifiers. Thus, using the cross-modality and cross-domain co-regularizer, the knowledge of RGB-D images in the source domain is transferred to the target domain to improve the target classifier. The results of the experiment show that the proposed method is one of the most effective ones.
Reinforcement learning (RL) constitutes an effective method of controlling dynamic systems without prior knowledge. One of the most important and difficult problems in RL is the improvement of data efficiency. Probabilistic inference for learning control (PILCO) is a state-of-the-art data-efficient framework that uses a Gaussian process to model dynamic systems. However, it only focuses on optimizing cumulative rewards and does not consider the accuracy of a dynamic model, which is an important factor for controller learning. To further improve the data efficiency of PILCO, we propose its active exploration version (AEPILCO) that utilizes information entropy to describe samples. In the policy evaluation stage, we incorporate an information entropy criterion into long-term sample prediction. Through the informative policy evaluation function, our algorithm obtains informative policy parameters in the policy improvement stage. Using the policy parameters in the actual execution produces an informative sample set; this is helpful in learning an accurate dynamic model. Thus, the AEPILCO algorithm improves data efficiency by learning an accurate dynamic model by actively selecting informative samples based on the information entropy criterion. We demonstrate the validity and efficiency of the proposed algorithm for several challenging controller problems involving a cart pole, a pendubot, a double pendulum, and a cart double pendulum. The AEPILCO algorithm can learn a controller using fewer trials compared to PILCO. This is verified through theoretical analysis and experimental results.
S. Zheng, Guohao Wu, Suoliang Zhang, Jie Su, Lei Liu, Fang Wang, Rui Zhao and Xiaobing Yan
The electronic structures of Hg-doped anatase TiO2 with different O vacancy concentrations were calculated using the first-principles based on the density functional theory. The calculated results show that the forbidden band widths of Hgdoped anatase TiO2 widened along with the increase of O vacancy concentration, which is responsible for the blue shift in the absorption edges. It can be deduced from the present study that the Hg-doped TiO2 samples prepared in the experimental research contain a certain quantity of O vacancies.
Jia-San Zheng, Jing-Nie, Ting-Ting Zhu, Hong-Ri Ruan, Xue-Wei and Rui-Wu
The value of neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule-1 (Kim-1), and liver-type fatty acid binding protein (L-FABP) was assessed in early diagnosis of gentamicin-induced acute kidney injury (AKI) in dogs.
Material and Methods
Subcutaneous gentamicin injection in 16 healthy adult beagles made the AKI model. Blood was sampled every 6 h to detect NGAL, Kim-1, L-FABP, and serum creatinine (SCr) concentrations. Kidney tissue of two dogs was taken before the injection, as soon as SCr was elevated (78 μmol/L), and when it had risen to 1.5 times the baseline, and haematoxylin-eosin staining and transmission electron microscopy (TEM) were used to observe changes.
NGAL, Kim-1, and SCr levels were significantly increased (P < 0.05) at 18, 30, and 78 h post injection, but L-FABP concentration was not associated with renal injury. At the earliest SCr elevation stage, findings were mild oedema, degeneration, and vacuolisation in renal tubular epithelial cells in pathology, and mild cytoplasmic and mitochondrial oedema in TEM. At this time point, NGAL and Kim-1 concentrations were significantly increased (P < 0.05), indicating that these two molecules biomark early kidney injury in dogs. Using receiver operating characteristic curve analysis, their warning levels were > 25.31 ng/mL and > 48.52 pg/mL.
Plasma NGAL and Kim-1 above warning levels are early indicators of gentamicin-induced AKI in dogs.
Shengdong Zhu, Pei Yu, Mingke Lei, Yanjie Tong, Lu Zheng, Rui Zhang, Jun Ji, Qiming Chen and Yuanxin Wu
Ionic liquid (IL) pretreatment of lignocellulosic materials has provided a new technical tool to improve lignocellulosic ethanol production. To evaluate the influence of the residual IL in the fermentable sugars from enzymatic hydrolysis of IL pretreatment of lignocellulosic materials on the subsequent ethanol fermentation, the toxicity of the IL 1-butyl-3-methylimidazolium chloride ([BMIM]Cl) to Saccharomyces cerevisiae AY93161 was investigated. Firstly, the morphological structure, budding and metabolic activity of Saccharomyces cerevisiae AY93161 at different [BMIM]Cl concentrations were observed under an optical microscope. The results show that its single cell morphology remained unchanged at all [BMIM]Cl concentrations, but its reproduction rate by budding and its metabolic activity decreased with the [BMIM]Cl concentration increasing. The half effective concentration (EC50) and the half inhibition concentration (IC50) of [BMIM]Cl to Saccharomyces cerevisiae AY93161 were then measured using solid and liquid suspension culture and their value were 0.53 and 0.39 g.L-1 respectively. Finally, the influence of [BMIM]Cl on ethanol production was investigated. The results indicate that the [BMIM]Cl inhibited the growth and ethanol production of Saccharomyces cerevisiae AY93161. This toxicity study provides useful basic data for further development in lignocellulosic ethanol production by using IL technology and it also enriches the IL toxicity data.