Few-shot learning in image classification aims to learn a classifier to classify images when only few training examples are available for each class. Recent work has achieved promising classification performance, where an image-level feature based measure is usually used. In this paper, we argue that a measure at such a level may not be effective enough in light of the scarcity of examples in few-shot learning. Instead, we think a local descriptor based image-to-class measure should be taken, inspired by its surprising success in the heydays of local invariant features. Specifically, building upon the recent episodic training mechanism, we propose a Deep Nearest Neighbor Neural Network (DN4 in short) and train it in an end-to-end manner. Its key difference from the literature is the replacement of the image-level feature based measure in the final layer by a local descriptor based image-to-class measure. This measure is conducted online via a k-nearest neighbor search over the deep local descriptors of convolutional feature maps. The proposed DN4 not only learns the optimal deep local descriptors for the image-to-class measure, but also utilizes the higher efficiency of such a measure in the case of example scarcity, thanks to the exchangeability of visual patterns across the images in the same class. Our work leads to a simple, effective, and computationally efficient framework for few-shot learning. Experimental study on benchmark datasets consistently shows its superiority over the related stateof-the-art, with the largest absolute improvement of 17% over the next best. The source code can be available from https://github.com/WenbinLee/DN4.git.

Few-shot learning aims to recognize new concepts from very few examples. However, most of the existing few-shot learning methods mainly concentrate on the first-order statistic of concept representation or a fixed metric on the relation between a sample and a concept. In this work, we propose a novel end-to-end deep architecture, named Covariance Metric Networks (CovaMNet). The CovaMNet is designed to exploit both the covariance representation and covariance metric based on the distribution consistency for the few-shot classification tasks. Specifically, we construct an embedded local covariance representation to extract the second-order statistic information of each concept and describe the underlying distribution of this concept. Upon the covariance representation, we further define a new deep covariance metric to measure the consistency of distributions between query samples and new concepts. Furthermore, we employ the episodic training mechanism to train the entire network in an end-to-end manner from scratch. Extensive experiments in two tasks, generic few-shot image classification and fine-grained fewshot image classification, demonstrate the superiority of the proposed CovaMNet. The source code can be available from https://github.com/WenbinLee/CovaMNet.git.

The discovery of IsoPs as products of nonenzymatic lipid peroxidation has opened up new areas of investigation regarding the role of free radicals in human physiology and pathophysiology. The quantification of IsoPs as markers of oxidative stress status appears to be an important advance in our ability to explore the role of free radicals in the pathogenesis of human disease. An important need in the field of free-radical medicine is information regarding the clinical pharmacology of antioxidant agents. Because of the evidence implicating free radicals in the pathogenesis of a number of human diseases, large clinical trials are planned or underway to assess whether antioxidants can either prevent the development or ameliorate the pathology of certain human disorders. However, data regarding the most effective doses and combination of antioxidant agents to use in these clinical trials is lacking. As mentioned previously, administration of antioxidants suppresses the formation of IsoPs, even in normal individuals. Thus, measurement of IsoPs may provide a valuable approach to define the clinical pharmacology of antioxidants. In addition to being markers of oxidative stress, several IsoPs possess potent biological activity. The availability of additional IsoPs in synthetic form should broaden our knowledge concerning the role of these molecules as mediators of oxidant stress. Despite the fact that considerable information has been obtained since the initial report of the discovery of IsoPs [6], much remains to be understood about these molecules. With continued research in this area, we believe that much new information will emerge that will open up additional important new areas for future investigation.

No abstract

The bipyridyl herbicide, diquat, is a potent prooxidant that generates superoxide anions through redox cycling in vivo. Exposure to elevated levels of this compound causes acute hepatic and renal toxicity as well as death in rodents. In the present study, we investigated whether melatonin, a free radical scavenger and antioxidant, could protect against diquat-induced hepatic and renal damage and whether the indole would improve survival of Kunming mice given a lethal dose of diquat. When mice were intraperitoneally (i.p.) given a single dose of diquat (50 mg/kg body weight), liver and kidney injuries were observed at 6 hr as indicated by elevated serum levels of both alanine aminotransferase (ALT) activity and blood urea nitrogen (BUN). In addition, lipid peroxidation levels in both liver and kidney showed significant increases as shown by elevated concentrations of F(2)-isoprostanes. The administration of melatonin (20 mg/kg) 30 min before the diquat injection resulted in a significant reduction in serum levels of ALT and BUN as well as hepatic and renal F(2)-isoprostanes levels. For the survival study, 75 mg/kg diquat was administered i.p. into mice to induce acute death. Without melatonin treatment, 10 of 23 (43.5%) mice died within 24 hr after diquat injection. Pretreatment with melatonin (20 mg/kg) 30 min prior to the injection of diquat and thereafter at 4-hr intervals until the end of the observation period (24 hr), reduced the death rate to two of 22 (9.1%) mice. Chi-squared test revealed a significant difference with P< or = 0.05. In conclusion, melatonin, a broad spectrum antioxidant, reduces hepatic and renal damage and lowers the death rate in diquat-treated mice.