Eleni Bougioukou, Nikolaos Toulgaridis, Maria Varsamou and Theodore Antonakopoulos:

Hardware Acceleration on Cloud Services: The use of Restricted Boltzmann Machines on Handwritten Digits Recognition

Advances in Science, Technology and Engineering Systems Journal, Vol. 3, No. 1, 483-495 (2018).

Abstract: Cloud computing allows users and enterprises to process their data in high performance servers, thus reducing the need for advanced hardware at the client side. Although local processing is viable in many cases, collecting data from multiple clients and processing them in a server gives the best possible performance in terms of processing rate. In this work, the implementation of a high performance cloud computing engine for recognizing handwritten digits is presented. The engine exploits the benefits of cloud and uses a powerful hardware accelerator in order to classify the images received concurrently from multiple clients. The accelerator implements a number of neural networks, operating in parallel, resulting to a processing rate of more than 10 MImages/sec.

If you need additional information concerning this paper, please contact either one of the authors or send an e-mail to: comes-sup@ece.upatras.gr