Share
Embed
Introduction to Google Cloud Platform CloudML with Qwiklabs
Talk conducted during Google Cloud Platform Next '18 Extended Cagayan de Oro at Department of Information and Communications Technology (DICT) Training Center, 54 T Chavez St, Cagayan de Oro.
Marc Anthony Reyes
September 15, 2018
Transcript
34 slides
-
Unstructured data accountsfor 90% of enterprise data* Cloud Machine Learning
helpyou make sense ofit *Source:IDC -
The point of ML is to make predictions Input Feature
PredictedValue odel M Google CloudPlatform Confidential &Proprietary 9 -
Tensorflow helps you “train”models Input Feature PredictedValue odel M TrueValue
Update model based onCost Cost Google CloudPlatform Confidential &Proprietary 10 -
Democratizing machinelearning AppDeveloper DataScientist CloudML Build custommodels Use/extend OSSSDK Scale,No-ops
Infrastructure Vision API SpeechAPI Use pre-builtmodels ML APIs TranslateAPI ML researcher Language API Google CloudPlatform Confidential &Proprietary 11 -
Beyond Tensorflow Scale of Compute Problem Accuracy CloudML(a) Deepnetworks TensorFlow
Processing Units(TPUs) Distributed No-ops Size ofdataset Size ofNN https://cloudplatform.googleblog.com/2016/05/Google-supercharges-machine-learning-tasks-with-custom- chip.html Vision API SpeechAPI ML APIs TranslateAPI Language API Google CloudPlatform Confidential &Proprietary 12 -
ML APIs are simply REST callsand can be made from
any language orframework sservice = build('speech', 'v1beta1', developerKey=APIKEY) response =sservice.speech().syncrecognize( body={ 'config': { 'encoding': 'LINEAR16', 'sampleRate': 16000 }, 'audio': { 'uri': 'gs://cloud-training-demos/vision/audio.raw' } }).execute() print response Data on CloudStorage Google CloudPlatform Confidential &Proprietary 13 -
Enterprise Predictive AnalyticsChallenges Data access to avariety of datasources. Develop
andbuild analytic models. Data preparation, exploration andvisualization. Deploy models and integrate them into businessprocesses and applications. High performance andscalability for both development and deployment. Perform platform,project and model management. Google CloudPlatform Confidential &Proprietary 14 -
Data Warehouse is theFoundation of SomethingBigger Data Warehouses/Lakes Machine Intelligence
Predictive + Prescriptive Analytics = Advanced Analytics Cloud On Premises Machine Learning APIs Train YourOwn Models Google CloudPlatform Confidential &Proprietary 15 -
Machine Learning UseCases •Predictive maintenance or condition monitoring •Warranty reserve
estimation •Propensity to buy •Demand forecasting •Process optimization Manufacturing • Predictive inventory planning • Recommendation engines • Upsell and cross-channelmarketing • Market segmentation andtargeting • Customer ROI and lifetime value Retail •Alerts and diagnostics from real-time patient data •Disease identification and risksatisfaction •Patient triage optimization •Proactive healthmanagement •Healthcare provider sentiment analysis Healthcare and LifeSciences •Aircraft scheduling •Dynamic pricing •Social media – consumer feedback and interaction analysis •Customer complaint resolution •Traffic patterns andcongestion management Travel andHospitality •Risk analytics andregulation •Customer Segmentation •Cross-selling and up-selling •Sales and marketing campaign management •Credit worthiness evaluation Financial Services • Power usage analytics • Seismic data processing • Carbon emissions andtrading • Customer-specific pricing • Smart grid management • Energy demand and supplyoptimization Energy, Feedstock andUtilities Google CloudPlatform Confidential &Proprietary 16 -
Why So Little Machine Learning Apps OutThere? Google CloudPlatform Confidential
&Proprietary 17 • Building and scaling machine learning infrastructureis hard • Operating production ML system is time consuming and expensive -
Building Smart Applications Today Google CloudPlatform Confidential &Proprietary 18 Technology
Operationalization Tooling Difficult toscale Many choices fordifferent use cases Using latest technology(e.g. DNN) is hard Complex data pipelines Managing ML infratakes away time from actually doing ML Many models tomanage Complex dev pipelinewith many combinations of tools/libraries Not fully interactive developer experience - collaboration/sharing ishard -
Introducing Cloud MachineLearning • Fully managedservice • Train using a
custom TensorFlowgraph for any ML usecases • Training at scale to shorten dev cycle • Automatically maximizepredictive accuracy with HyperTune • Batch and online predictions, atscale • Integrated Datalabexperience Google CloudPlatform Confidential &Proprietary 19 -
Powerful Machine LearningAlgorithm Google CloudPlatform Confidential &Proprietary 21 • Convolutional
Neural Network for image classification • Recursive Neural network fortext sentiment analysis • Linear regression at scale to predict consumer action (purchase prediction, churnanalysis) • And unlimited variety of algorithms you can build using TensorFlow -
Automatically tune your model withHyperTune • Automatic hyperparametertuning service •
Build better performingmodels faster and save many hours of manual tuning • Google-developed search algorithm efficiently finds better hyperparameters for your model/dataset Objective Want to findthis Not these Google CloudPlatform Confidential &Proprietary 22 -
Integrated with GCPProducts Google CloudPlatform Confidential &Proprietary 23 • Access
data that is stored in GCSor BigQuery • Save trained models toGCS • Preprocess largest datasets (TB) usingDataflow • Orchestrate ML workflowas a Dataflow pipeline • Analyze data and interactively develop ML models in Datalab • AutoML for premade and customized ML models curated for your specific ML projects -
Fully Managed Machine LearningServices Google CloudPlatform Confidential &Proprietary 24 •
Scalable and distributed training infrastructure foryour largest data sets • Scalable prediction infrastructure thatcan serve very large traffic • Managed no-ops infrastructure handles provisioning, scaling, and monitoring so that you can focus on building your models instead of handlingclusters -
Transform Data intoActions Databases Storage Mobileapps Sensorsand devices Webapps Relational
Key-value Document SQL Widecolumn Object Data Preparation& Processing Data preparation Stream processing Batch processing Analytics Federated query Datacatalog Exploration& Collaboration Data exploration Data visualization Developers Datascientists Business analysts Advanced Analytics & Intelligence Development environment for Machine Learning Pre-Trained Machine Learning models Data Ingestion Logs Messaging Google CloudPlatform Confidential &Proprietary 29 -
Transform Data intoActions Data Preparation& Processing CloudDataflow CloudDataproc Exploration& Collaboration
Google BigQuery CloudDatalab Google Analytics360 CloudDataproc Mobileapps Sensorsand devices Webapps Developers Datascientists Business analysts DataIngestion CloudPub/Sub AppEngine Databases / Storage CloudSQL CloudBigtable Cloud Datastore CloudStorage Analytics GoogleBigQuery Google Analytics360 CloudDataproc GoogleDrive Advanced Analytics & Intelligence CloudMachine Learning TranslateAPI VisionAPI SpeechAPI Google CloudPlatform Confidential &Proprietary 30 -
Use Your Own Data to Train Models BETA GA GA
CloudDatalab BETA Cloud MachineLearning CloudStorage GoogleBigQuery Develop/Model/Test Google CloudPlatform Confidential &Proprietary 31 -
HTTPrequest Use your own data to train models DataStorage Trainingflow
Prediction flow Local training Download Mobile predictio n Batch Online Pre-Processing Training Prediction Tooling Datalab Datalab Tooling Upload HostedModel Google CloudPlatform Confidential &Proprietary 32 -
Automatically categorize, and automatically extract value 1 Identify categorizations that
provide value, categories you’re already evaluating for by handtoday 2 Capture thousands of examples of correct evaluations for that categorization, and use them to train an ML model 3 Evaluate the model by applying it against additional manually categorized data,correct and tune 4 Google CloudPlatform Confidential &Proprietary 33 Machine Intelligence is Already Making a Huge Difference and There are Many, Many MoreOpportunities -
Codelabs this Afternoon Google CloudPlatform Confidential &Proprietary 36 • Cloud
ML Engine: Qwik Start • Cloud Natural Language API: Qwik Start -
Awesome Stuff for Participants Google CloudPlatform Confidential &Proprietary 37 •
One-month free access to Qwiklabs with 150 credits for free platform usage. • One-month free access to GCP Introductory Course on Coursera. -
Access These Links for the Qwiklab Activity Google CloudPlatform Confidential
&Proprietary 38 • Baseline: Data, ML, AI Quest • Cloud ML Engine: Qwik Start (http://bit.ly/GCP18CDOCloudML) • Cloud Natural Language API: Qwik Start (http://bit.ly/GCP18CDONL) • This afternoon, we only will be taking Cloud ML Engine: Qwik Start and Cloud Natural Language API: Qwik Start