Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/72827
Title: Distributed machine learning on public clouds
Authors: Tran, Manh Tu
Keywords: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Issue Date: 2017
Abstract: Machine learning (ML) is prevalent in today’s world. Starting from the need to improve artificial intelligent, scientist attempted to have machines learning from data. Today, machine learning has been used in a wide range of applications from sciences such as bioinformatics, brain-machine interfaces, classifying DNA sequences to smart ecosystems with computer vision, natural language processing, speech recognition and many more. Due to its ease of adaptation, more and more people are starting to use its for their own use cases. However, they will need to have some good hardware as a start which not everyone can afford. Public cloud service providers are providing the infrastructure for machine learning at a more affordable cost such as EC2 from Amazon, Azure from Microsoft, GCE from Google. In the past when use cases of machine learning was limited and the amount of data available is small, machine learning can be done using a single machine. That is not the case in today’s world where exabytes of data are collected in a single day. Distributed machine learning arises as the solution for large scale machine learning. However, in distributed settings, machine learning does not scale linearly as more nodes are used. This project aims to construct predictive models to predict the optimal distributed settings for a sample dataset.
URI: http://hdl.handle.net/10356/72827
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
amended_report.pdf
  Restricted Access
Amended FYP Report1.04 MBAdobe PDFView/Open

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.