Microsoft Azure ML Machine Learning as a Service

Microsoft Azure ML: Machine Learning as a Service Dmitry Petukhov #Moscow. Data. Fest

Challenge Local PC some library Yet Another Runtime Execution Engine Python Runtime ML Framework scikit learn Hybrid Model Python / R tools Cluster (on-premises/cloud) MLlib Mahout Spark Map. Reduce YARN / Apache Mesos YARN HDFS / S 3 HDFS ML as a Service (cloud) Python / R tools Python / R on Spark Resource Management Local OS YARN / Apache Mesos Storage Local Disc HDFS / S 3 Dark Magic… Distributed FS

Intro <- function() { Hello Data Fest! I need your help } Agenda Learn <- function() { Azure ML Overview # +Hello Azure ML Demo Data Science Workflow vs Azure ML } Code <- function() { ML Skills Cluster Analysis # Demo 1 Twitter sentiment analysis # Demo 2 } Coffee <- function() { Q&A Contacts }

Azure Machine Learning. Introduction Dmitry Petukhov, Hello Data Fest! Software Architect + Developer, Microsoft Certified Professional (C#), Big Data Enthusiast && Coffee Addict Researcher & Developer @ Open. Way

Azure Machine Learning. Overview Data Science is far too complex today Math Computer Science Domain Guiding Principles Reduce complexity to broaden participation No software to install, only web browser; Possibility to develop without writing line of code; Easy deployment and usage using restfull API; Easy collaboration on Azure ML projects; Visual composition with end 2 end support for Data Science workflow; Extensible, support for R OSS. Reference: Tech. Ed 2014 Conference

Azure Machine Learning. Overview Azure Machine Learning Data Cloud storage Azure Storage Azure Table Hive etc. Local storage Upload data from PC… Business problem ML Studio (Web IDE) Model Consumers ML Web Services API (REST API Services) Excel Manage Workspace: Experiments Datasets Trained models Notebooks Access settings Modeling Azure Marketplace (Applications store) Azure ML Gallery (community) Deployment API Business Apps Business value Reference: Tech. Ed 2014 Conference

Azure Machine Learning. Overview Step 1. Get $200 credit Sign up for Azure free trial. Demo #0: Hello Azure ML! Step 2. Get access to Azure Portal Step 3. Create Azure ML Workspace Step 4. Go to Azure ML Studio & create ML Experiment Step 5. Publish result

Azure Machine Learning. Azure ML Flow Supervised Learning Flow Part #1

Azure Machine Learning. Azure ML Flow Supervised Learning Flow Part #2 Source

Azure Machine Learning. Azure ML Flow Source: Azure ML Cheat Sheet

Azure Machine Learning. Demo #1: k-means clustering aims to partition the n observations into k (≤ n) sets S = {S 1, S 2, …, Sk} so as to minimize the within-cluster sum of squares (WCSS). ML Skills Cluster Analysis where (x 1, x 2, …, xn) – observations, μi is the mean of points in Si. Source: Wikipedia

Azure Machine Learning. Demo TD-IDF, short for term frequency–inverse document frequency, is a numerical statistic that is intended to reflect how important a word is to a document in a collection or corpus. Demo #2: Twitter sentiment analysis Source: Wikipedia

Azure Machine Learning. Conclusion Legislative restrictions International & local Azure platform restrictions Max storage volume per account, etc. Restrictions Azure ML service restrictions Data Max dataset volume: 10 Gb Vector size limitation: 2^64 Throttled policy 20 concurrent request per endpoint Max endpoints count: 10 K Black box No debug No Scala, C++, C# No your own right algorithms

Azure Machine Learning. Conclusion R (quickstart) Support R models & scripts Python (quickstart) Support Python scripts Jupyter Notebooks in Azure ML Studio Killer Features Publishing REST API & real-time mode vs batch-mode Azure ML Gallery Share for community Azure Marketplace Saa. S store In-the-box integration with… Hive, Azure Storage, Excel, Cortana Analytics Stack Free Start & it’s child age

Azure Machine Learning. Conclusion Data Science still too complex today Math Computer Science Domain Nothing has changed Reduce complexity to broaden participation No software to install, only web browser; Possibility to develop without writing line of code; Easy deployment and usage using restfull API; Easy collaboration on Azure ML projects; Visual composition with end 2 end support for Data Science workflow; Extensible, support for R OSS.

Azure Machine Learning. Conclusion Start from azure. com/ml Microsoft Machine Learning Blog References Azure ML documentation + free online course, videos & books Microsoft Research: Azure for Researchers

Thank you! © 2015 Dmitry Petukhov All rights reserved. Microsoft Azure and other product names are or may be registered trademarks and/or trademarks in the U. S. and/or other countries.

Azure ML: Machine Learning as a Service Q&A Now or later (send on d. petukhov@outlook. com) Stay connected Facebook: @code. zombi Linked. In: @dpetukhov Habr: @codezombie All contacts… Read my tech code instinct blog Download presentation from http: //0 xcode. in/moscow-data-fest or