DISTRIBUTION OF THE NEURAL NETWORK BETWEEN MOBILE DEVICE AND CLOUD INFRASTRUCTURE SERVICES
Journal Title: Современные информационные технологии и ИТ-образование - Year 2018, Vol 14, Issue 4
Abstract
Neural networks become the only way to solve problems in some areas. Such tasks as recognition of images, sounds, classification require serious processor power and memory for training and functioning of the network. Modern mobile devices have quite good characteristics for primary layers of deep neural networks, but there are not enough resources for whole network. Since neural networks for mobile devices are trained separately on external resources, a method of distributed work of a neural network with vertical distribution over sets of layers with synchronization of training data was developed. The model is divided after saving its state, all layers on the mobile device are converted to the format for the mobile framework and synchronized with the device after training on a distributed platform. Variables and coefficients are formed separately, which allows to significantly reduce the size of the neural network data file uploaded to the device. An algorithm for automatic selection of a neural network separation point was proposed. It based on the data amount transferred between the layers and the load on the mobile device resources. The approach allows to use full-size deep neural networks with a mobile device. Performance experiment showed possibility of obtains an acceptable response even with an unstable communication channel without overloading communication channels and device resources.
Authors and Affiliations
Yuri Ushakov, Petr Polezhaev Petr, Aleksandr Shukhman, Margarita Ushakova
Keywords
Related Articles
- EP ID EP523569
- DOI 10.25559/SITITO.14.201804.903-910
- Views 99
- Downloads 0
NEURAL NETWORK METHOD OF RESTORING AN INITIAL PROFILE OF THE SHOCK WAVE
In this paper, we apply neural network modeling to solve the inverse problem of mathematical physics with a system of nonlinear partial differential equations of hyperbolic type. In the problem, the initial conditions ar...
DEVELOPMENT OF THE FIRST PART OF THE AUTOMATED SYSTEM FOR TEACHING COMPUTATIONAL COMPLEXITY OF ALGORITHMS COURSE
Previous articles [1-2] focus on researching the software for developing the automated teaching system (ATS) “Algorithm complexity analysis”. This article addresses issues concerning overall learning process features and...
THE METHOD OF DEVELOPING AN INTELLIGENT AUTOMATED SUBSYSTEM FOR ASSESSING THE LEVEL OF COMPETENCE FORMATION
In the article the question of an estimation of reliability of level of formation of the competences on the basis of indistinct production cognitive maps is considered. The paper considers the issue of assessing the reli...
RESOURCE VOLUME OPTIMIZATION OF LANGUAGE IDENTIFIERS FOR THE ACCOUNT OF METHODS AND IDENTIFICATION ALGORITHMS
This article is a continuation of the author's series of publications on the topic of text’s language identification. Here is considered the possibility of optimizing the resource capacity of programs and systems for lan...
NETWORK ORGANIZATION AND SYSTEMIC PRINCIPLES OF GOVERNANCE IN RELATIONSHIPS BETWEEN INNOVATIVE BUSINESS AND THE STATE
Network organization and systemic principles of governance in relationships between innovative business and the state are examined on the basis of the innovation clusters formation. The focus is on the formation of innov...