National Research University – Higher School of Economics

About: National Research University – Higher School of Economics is a education organization based out in Moscow, Russia. It is known for research contribution in the topics: Population & Computer science. The organization has 12873 authors who have published 23376 publications receiving 256396 citations.

OpenVINO Deep Learning Workbench: Comprehensive Analysis and Tuning of Neural Networks Inference

Abstract: A task of maximizing deep learning neural networks performance is a challenging and actual goal of modern hardware and software development. Regardless the huge variety of optimization techniques and emerging dedicated hardware platforms, the process of tuning the performance of the neural network is hard. It requires configuring dozens of hyper parameters of optimization algorithms, selecting appropriate metrics, benchmarking the intermediate solutions to choose the best method, platform etc. Moreover, it is required to setup the hardware for the specific inference target. This paper introduces a sophisticated software solution (Deep Learning Workbench) that provides interactive user interface, simplified process of 8-bit quantization, speeding up convolutional operations using the Winograds minimal filtering algorithms, measuring accuracy of the resulting model. The proposed software is built over the open source OpenVINO framework and supports huge range of modern deep learning models.

Latarcins: versatile spider venom peptides

Abstract: Arthropod venoms feature the presence of cytolytic peptides believed to act synergetically with neurotoxins to paralyze prey or deter aggressors. Many of them are linear, i.e., lack disulfide bonds. When isolated from the venom, or obtained by other means, these peptides exhibit common properties. They are cationic; being mostly disordered in aqueous solution, assume amphiphilic α-helical structure in contact with lipid membranes; and exhibit general cytotoxicity, including antifungal, antimicrobial, hemolytic, and anticancer activities. To suit the pharmacological needs, the activity spectrum of these peptides should be modified by rational engineering. As an example, we provide a detailed review on latarcins (Ltc), linear cytolytic peptides from Lachesana tarabaevi spider venom. Diverse experimental and computational techniques were used to investigate the spatial structure of Ltc in membrane-mimicking environments and their effects on model lipid bilayers. The antibacterial activity of Ltc was studied against a panel of Gram-negative and Gram-positive bacteria. In addition, the action of Ltc on erythrocytes and cancer cells was investigated in detail with confocal laser scanning microscopy. In the present review, we give a critical account of the progress in the research of Ltc. We explore the relationship between Ltc structure and their biological activity and derive molecular characteristics, which can be used for optimization of other linear peptides. Current applications of Ltc and prospective use of similar membrane-active peptides are outlined.