A Few-Shot Learning-Based Siamese Capsule Network for Intrusion Detection with Imbalanced Training Data.

Network intrusion detection remains one of the major challenges in cybersecurity. In recent years, many machine-learning-based methods have been designed to capture the dynamic and complex intrusion patterns to improve the performance of intrusion detection systems. However, two issues, including imbalanced training data and new unknown attacks, still hinder the development of a reliable network intrusion detection system. In this paper, we propose a novel few-shot learning-based Siamese capsule network to tackle the scarcity of abnormal network traffic training data and enhance the detection of unknown attacks. In specific, the well-designed deep learning network excels at capturing dynamic relationships across traffic features. In addition, an unsupervised subtype sampling scheme is seamlessly integrated with the Siamese network to improve the detection of network intrusion attacks under the circumstance of imbalanced training data. Experimental results have demonstrated that the metric learning framework is more suitable to extract subtle and distinctive features to identify both known and unknown attacks after the sampling scheme compared to other supervised learning methods. Compared to the state-of-the-art methods, our proposed method achieves superior performance to effectively detect both types of attacks.

Topiramate reduced sweat secretion and aquaporin-5 expression in sweat glands of mice.

Decreased sweat secretion is a primary side effect of topiramate in pediatric patients, but the mechanism underlying this effect remains unclear. This study aimed to better understand how topiramate decreases sweat secretion by examining its effect on the expression of carbonic anhydrase (CA) II and aquaporin-5 (AQP5), total CA activity, as well as on tissue morphology of sweat glands in mice. Both developing and mature mice were treated with a low (20 mg/kg/day) and high dose (80 mg/kg/day) of topiramate for 4 weeks. Sweat secretion was investigated by an established technique of examining mold impressions of hind paws. CA II and AQP5 expression levels were determined by immunofluorescence and immunoblotting and CA activity by a colorimetric assay. In mature mice, topiramate treatment decreased the number of pilocarpine reactive sweat glands from baseline in both the low and high dose groups by 83% and 75%, respectively. A similar decrease was seen in developing mice. Mature mice with reactive sweat glands that declined more than 25% compared to baseline were defined as anhidrotic mice. These mice did not differ from controls in average secretory coil diameter, CA II expression and CA activity. In contrast, anhidrotic mice did show a reduction in membrane AQP5 expression in sweat glands after topiramate delivery. Thus, sweat secretion and membrane AQP5 expression in mouse sweat glands decreased following topiramate administration. These results suggest dysregulation of AQP5 may be involved in topiramate-induced hypohidrosis and topiramate may serve as a novel therapy for hyperhidrosis.

Postnatal expression and denervation induced up-regulation of aquaporin-5 protein in rat sweat gland.

The evolution of aquaporin-5 (AQP5) expression during postnatal development has not been defined in the sweat gland. Previous studies have suggested that AQP isoforms in several peripheral targets are regulated by a neural mechanism. We have examined, in rat sweat glands, the expression of AQP5 during postnatal development and the effects of denervation on AQP5 expression. Both AQP5 mRNA and protein begin to be expressed at postnatal day 10, before sweat-secretory responsiveness first appears; this expression coincides with the occurrence of vasoactive intestinal peptide (VIP) immunoreactivity. Early noradrenergic and later cholinergic interaction between sweat glands and their innervation are disrupted by neonatal chemical sympathectomy or postnatal severance of the sciatic nerve. Examination of such denervated developing rats has shown that secretory responsiveness fails to arise later in the adults, and AQP5 immunostaining increases in the denervated glands, whereas gland morphogenesis and the occurrence of AQP5 expression proceed normally. Immunobloting has revealed an increase of AQP5 abundance after the denervated mature glands lose their secretory ability. These findings suggest that AQP5 protein is necessary for sweat secretion, and that the expression of AQP5 in rat sweat glands is independent of sympathetic innervation. Our data also indicate that factor(s) regulating the normal morphological development of sweat gland might be responsible for controlling AQP5 expression.