A Lightweight Cloud Execution Stack for Neural Network Simulation

A Lightweight Cloud Execution Stack for Neural Network Simulation

Abstract

This paper presents an execution stack for neural network simulation using Cloud container orchestration and microservices. User (or other systems) can employ it by simple RESTful service calls. This service oriented approach allows easy and user-friendly importing, training and evaluating of arbitrary neural network models. This work is influenced by N2Sky, a framework for the exchange of neural network specific knowledge and is based on ViNNSL, the Vienna Neural Network Specification Language, a domain specific neural network modelling language. The presented execution stack runs on many common cloud platforms. It is scalable and each component is extensible and interchangeable.

Grafik Top
Authors
  • Nussbaum, Benjamin
  • Schikuta, Erich
Grafik Top
Shortfacts
Category
Paper in Conference Proceedings or in Workshop Proceedings (Paper)
Event Title
25th International Conference on Neural Information Processing (ICONIP 2018)
Divisions
Workflow Systems and Technology
Subjects
Webentwicklung, Webanwendungen
Kuenstliche Intelligenz
Event Location
Siem Reap, Cambodia
Event Type
Conference
Event Dates
13-16 Dec 2018
Date
December 2018
Export
Grafik Top