Democratizing Artificial Intelligence
Download as PPTX, PDF•1 like•643 views
This document summarizes the evolution of machine learning tools in the cloud from 2015 to 2017. It describes how in 2015, major cloud providers like Azure, Amazon, and Google launched early machine learning services. From 2015 to 2016, these providers also released popular deep learning frameworks like TensorFlow as open source. During this time period, the providers began offering deep learning models and GPU computing as cloud services. The document argues that these developments have helped democratize artificial intelligence and machine learning.
Democratizing Artificial Intelligence
- 1. Democratizing Artificial Intelligence or Machine Learning for Elves 80 LvL #GlobalAzure Dmitry Petukhov, ML/DS Preacher, Machine Intelligence Researcher @ OpenWay && Coffee Addicted
- 2. Web Social Network s Science Healthcare Finance Telecom Retail Logistic Security Electronics Proof: https://www.kaggle.com/wiki/DataScienceUseCases Data Science Applications Data Science People
- 3. ML runtime as a Service • Feb 2015: Azure ML • Apr 2015: Amazon ML • Oct 2015: Google Cloud ML Engine Deep Learning algorithms in Open Source • Dec 2015: The Microsoft Cognitive Toolkit (CNTK) • Nov 2015: Microsoft Distributed Machine Learning Toolkit • Nov 2015: TensorFlow (Google) • May 2016: Amazon DSSTNE Deep Learning models as a Services • Microsoft Cognitive Services • Amazon Rekognition • Google APIs: Natural Language, Speech API, Translation API, etc. GPU on demand • Sep 2016: private preview in Azure • Oct 2017: Amazon update its GPU-instances • March 2017: Google GPU features 2015 2016 2017 Machine Learning Tools Evolution in Cloud
- 6. Twitter Sentiment Analysis with Azure Machine Learning 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 #Word2Vec #LDA #NLP Topic Model N-gram #GlobalAzure
- 7. What? Twitter Sentiment Analysis in real-time Why? Банк Инструменты атаки Дата начала атаки Информационный повод Сбербанк Массовая смс-рассылка, публикации в блогах, соцсетях, СМИ декабрь 2014 Введение санкций США и ЕС в отношении государственных банков РФ Банки Урала Массовая смс-рассылка, публикации в блогах, листовки сентябрь 2014 Отзыв лицензии у «Банк24.ру» Возрождение Массовая смс-рассылка октябрь 2013 Отзыв лицензии у банка «Пушкино» Банки Татарстана Массовая смс-рассылка, рассылка в мессенджерах, публикации в блогах, соцсетях и СМИ декабрь 2016 Финансовые проблемы банка «Пересвет» и группы Татфондбанка Source: FutureBanking…a total of more than 15 cases since 2013 Test It! #GlobalAzure
- 8. Near Real-time Predictive Pipeline Batch-mode Analytics Pipeline Twitter Semantic Analysis Architecture REST API h(θ0, θn) New Tweets Processing Azure Workers New Tweets Service Bus Azure Event Hub Tweets Stream Azure Stream Analytic Visualize result Power BI New Tweets Twitter Streaming API Semantic prediction API Azure ML Web Services Store Tweets Azure Table Data Sources Ingest Prepare ConsumeStream/Analyze 1 2 3 3 4 5 6 7 8 Calculate ML-models Azure Machine Learning Real-time processing Batch processingData flow Sentiment140 Dataset Azure Storage 0
- 9. Data Azure Machine Learning Consumers Cloud storage Business problem Modeling Business valueDeployment Model REST API Manage Local storage REST API Reference: Microsoft Data Amp 2017 ML Studio Web IDE for ML ML Web Services ML-model publication Azure Marketplace Monetize your model Data Model
- 10. References: 1. Source code on GitHub 2. Twitter Sentiment Experiment on Cortana Intelligence Gallery 3. Free access to Azure. How to? 4. Microsoft Data Amp 2017 Advanced references: 1. Microsoft Machine Learning Blog 2. Azure ML documentation 3. Free MOOCs on MVA and EdX 4. Microsoft Azure Russia Group in Facebook
- 11. © 2017, Dmitry Petukhov. CC BY-SA 4.0 license. Microsoft and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. Data Science must win!
- 12. Q&A Now or later (use contacts below) Ping me Habr: @codezombie All contacts: http://0xCode.in/author Download presentation from http://0xcode.in/2017/global-azure-bootcamp or
Editor's Notes
- #5: Cutting Edge 2015: ML Runtime as a Service Feb 2015: Azure ML Apr 2016: Amazon ML Oct 2015: Google Cloud ML Engine 2015-2016: Deep Learning algorithms in Open Source Dec 2015: The Microsoft Cognitive Toolkit (CNTK) Nov 2015: Microsoft Distributed Machine Learning Toolkit http://www.dmtk.io Nov 2015: TensorFlow https://github.com/tensorflow/tensorflow/ May 2016: Amazon DSSTNE https://github.com/amznlabs/amazon-dsstne 2016: Cognitive Services Microsoft Cognitive Services Amazon Rekognition Google APIs: Natural Language, Speech API, Translation API, etc. 2016-2017: GPU on demand Sep 2016: private preview in Azure Oct 2017: Amazon update its GPU-instances March 2017: Google
- #6: Source: https://blogs.technet.microsoft.com/dataplatforminsider/2017/04/19/delivering-ai-with-data-the-next-generation-of-microsofts-data-platform/
- #7: Microsoft R Open and Microsoft R Server R Server for Azure HDInsight #Hadoop #distributed Data Science VM Azure Machine Learning Azure HDInsight SQL Server R Services Excel / Power BI Python/R Tools for Visual Studio The Microsoft Cognitive Toolkit (CNTK) Microsoft Distributed Machine Learning Toolkit Cognitive Services to be continued…
- #8: Анализ тональности текста (sentiment analysis) – приложение методов обработки естественного языка (natural language processing, NLP), в частности, классификации, целью которой является извлечение из текста эмоционального содержания.
- #9: С 2013 года было зафиксировано 17 подтвержденных самими банками крупных информационных атак, в результате которых совокупные оттоки средств клиентов превысили 1,5 трлн рублей. Было зафиксировано две масштабных атаки на региональные финансовые системы (Урал, 2013 г., Татарстан, 2016 г.). По статистике в течение недели после начала подобных атак клиенты буквально выносили от 3% до 25% всех банковских пассивов на десятки миллиардов рублей. В итоге около четверти атакованных банков не пережили данные информационные кризисы и лишились лицензий или начали процедуру санации. Source: http://futurebanking.ru/post/3305
- #13: (c) 2017, Dmitry Petukhov. CC BY-SA 4.0 license.
- #14: Source: http://0xcode.in/2017/global-azure-bootcamp