Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Effective_MLOps_on_AWS.pdf

 Effective_MLOps_on_AWS.pdf

[First Tokyo WiMLDS Meeting]
by Tokyo Women in Machine Learning & Data Science

We are thrilled to announce the first event of Tokyo’s chapter of Women in Machine Learning and Data Science hosted and sponsored by Amazon Web Sevices Japan. All genders welcomed.

■ Venue  Meguro Central Square :3-1-1 Kamiosaki, Shinagawa-ku, Tokyo
■ Reception Meguro Central Square 17F
---

Title: Effective MLOps on AWS Cloud
Speaker: Shoko Utsunomiya. Machine Learning Solutions Architect at Amazon Web Services Japan.

Abstract:
In the operation of machine learning, there are various issues such as acquisition and annotation of high quality data, quick construction of training environment, preparation of elastic compute resources such as GPUs against demand fluctuation of calculation resource, and reduction of operation load of machine learning workflow.

Amazon Web Services (AWS) offers Amazon SageMaker, a fully managed machine learning platform service for solving these issues, and is used by customers of various sizes and stages. By removing Undifferentiated Heavy Lifting in a machine learning environment efficiently using managed services, you can focus on the more important differentiation issue, the data science tasks.

In this session, I will present the principles and best practices for machine learning in the cloud based on the past customer experiences. I will also introduce issues common to various use cases and their solutions, such as handling of large-scale data, transitioning research issues to actual operation, and launching of quick services.

Shoko Utsunomiya

May 18, 2019
Tweet

More Decks by Shoko Utsunomiya

Other Decks in Technology

Transcript

  1. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Effective MLOps on AWS Cloud Amazon Web Services Japan K.K. Machine Learning Solutions architect Shoko Utsunomiya
  2. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark About me Shoko Utsunomiya, Ph.D • Machine Learning Solutions Architect • Providing solutions of ML on AWS • My favorite AWS service • Amazon SageMaker • Worked for Automotive OEM as engineer of Autonomous driving system • Associate Professor of National Institute of Informatics (2013-2017)
  3. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark The mission and vision of Amazon is Earth's most customer-centric company
  4. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark ML @ AWS OUR MISSION Machine learning to every developer and data scientist
  5. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark "*.-VTF DBTFTJOWBSJPVTJOEVTUSJFT • $POUFOUTHFOFSBUJPO • 1SPNPUJPO NBSLFUJOH • $PQZSJHIU JOGSJOHFNFOU JOTQFDUJPO • "VUPUBHHJOHPG DPOUFOUT • "VUPNBUJDDBQUJPO DSFBUJPO • 'BJMVSFEFUFDUJPOPG QSPEVDUJPOMJOF • 7JTVBMJOTQFDUJPOPG GJOJTIFEQSPEVDUT • *NQSPWFDVTUPNFS TFSWJDFUISPVHIWPJDF TFSWJDFBOEDIBUCPU • $BMMDFOUFS PQUJNJ[BUJPO • 4FMGESJWJOHUFDIOJRVF • &DPNNFSDF • 1SPEVDU SFDPNNFOEBUJPO • 1FSTPOBMJ[F • $SFEJUFWBMVBUJPO • "EWFSUJTJOH NBSLFUJOH • 6OBUUFOEFETUPSF • 1BUJFOUTIFBMUI DPOEJUJPOCBTFE PODMJOJDBMEBUB • 1SFEJDUIPTQJUBM TUBZBOESF IPTQJUBMJ[BUJPO • %SVHEJTDPWFSZ • .FEJDBMJNBHJOH EJBHOPTJT • &BSMZEJBHOPTJT • &YBNJOBUJPO • 'SBVEEFUFDUJPO • .BSLFUJOH'VUVSF 'PSFDBTU • *OWFTUNFOUQPSUGPMJP NBOBHFNFOU • /FXT"OBMZTJT • (FPTQBUJBMJNBHFBOBMZTJT • $IBUCPUQFSTPOBMBEWJTPS .FEJB FOUFSUBJONFOU %JTUSJCVUJPO 3FUBJM )FBMUIDBSF -JGFTDJFODF 'JOBODF 4FSWJDF USBOTBDUJPO .BOVGBDUVSF "VUPNPUJWF
  6. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Tens of thousands of customers running Machine Learning on AWS
  7. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Confidential and Trademark Why do so many customers run machine learning workloads on AWS ?
  8. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Continuous ML development Data acquisition Pre- processing Model development Model evaluation Model translation Deployment Monitoring, evaluation
  9. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Continuous ML development Data acquisition Pre- processing Model development Model evaluation Model translation Deployment Monitoring, evaluation Very important to iterate this cycle as fast as possible to improve your AI...
  10. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Continuous ML development Data acquisition Pre- processing Model development Model evaluation Model translation Deployment Monitoring, evaluation Where should you spend your time?
  11. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark 8IBUJT.-0QT • A compound of Machine Learning and Operations • Practice for collaboration and communication between Data Scientists and Operations professionals to help manage production ML lifecycle • Looks to increase automation and improve the quality of production ML https://en.wikipedia.org/wiki/MLOps
  12. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Important things of MLOps in the cloud • Focus on business value • Not try too hard by yourself • Continuous improvement of ML models • Effective communication between Research and Operations • Rapid cycle from research to operation • Simple data flow • Remove undifferentiated heavy lifting
  13. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Remove undifferentiated heavy lifting
  14. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark 5PGPDVTPOZPVSCVTJOFTTWBMVF QSPEVDU 3FTFBSDI JEFB 1SPEVDU
  15. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark In fact... 3FTFBSDI JEFB )FBWZ MJGUJOH 1SPEVDU “Undifferentiated Heavy Lifting”
  16. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark “Undifferentiated Heavy Lifting” in Machine Learning &OWJSPONFOU • &TUJNBUFESFTPVSDFTOFFEFEBOEEFDJTJPOUPCVZ • 5JNFBOEFGGPSUUPQSPDVSFJOUIFGJSTUQMBDF • -PUTPGFGGPSUUPNBJOUBJOVOJGPSNFOWJSPONFOUUPEFWFMPQFST • 'SBNFXPSLJOTUBMMBUJPO WFSTJPODPOUSPM $POTUSVDUJPOPGPQUJNBM.-FOWJSPONFOUT • .BJOUBJOBOEVQEBUFPQUJNBM$16(16FOWJSPONFOUGPSUSBJOJOH • 4DBMBCMFSFTPVSDFTBOE%JTUSJCVUFE$PNNVOJDBUJPO&OWJSPONFOUGPS%JTUSJCVUFE -FBSOJOH 0QFSBUJPO • )PTUJOHFOWJSPONFOUQSFQBSBUJPOBOENPEFMIPTUJOH • 3FRVJSFEEJGGFSFOUTLJMMTFUTUIBOQVSFEBUBTDJFODFJTSFRVJSFE Photo by Victor Freitas on Unsplash
  17. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Remove undifferentiated heavy lifting to focus on your business Research idea Product
  18. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark 5 benefits of Machine Learning on AWS Cloud Use as much CPU / GPU as needed when needed System extensibility Improved agility for model building and service deployment Pay-as-you-go Cost optimization A lot of AWS Service / function Deploy worldwide instantly
  19. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark AWS Machine Learning Stack App developers with little knowledge of ML ML developers and data scientists ML researchers and academics Amazon SageMaker Ground Truth Algorithms Notebooks Marketplace Unsupervised Learning Supervised Learning Reinforcement Learning Optimization (Neo) Training Hosting Deployment Frameworks Interfaces Infrastructure Amazon Rekognition Image Amazon Polly Amazon Transcribe Amazon Translate Amazon Comprehend Comprehend medical Amazon Lex Amazon Rekognition Video Vision Speech Language Chatbots Amazon Forecast Forecasting Amazon Textract Amazon Personalize Recomme ndations Amazon EC2 P3 & P3DN Amazon EC2 C5 FPGAs AWS Greengrass Amazon Elastic Inference Amazon Inferentia Labeling Model development Training Hosting ML SERVICES ML FRAMEWORKS & INFRASTRUCTURE AI SERVICES New New New New New New New New New New New
  20. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark AWS Machine Learning App developers with little knowledge of ML Amazon Rekognition Image Amazon Polly Amazon Transcribe Amazon Translate Amazon Comprehend Comprehend medical Amazon Lex Amazon Rekognition Video Vision Speech Language Chatbots Amazon Forecast Forecasting Amazon Textract Amazon Personalize Recomme ndations AI SERVICES New New New New . : . . . Time series analysis Forecast Recommendation Personalize
  21. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark AWS Machine Learning ML developers and data scientists ML researchers and academics Amazon SageMaker Ground Truth Algorithms Notebooks Marketplace Unsupervised Learning Supervised Learning Reinforcement Learning Optimization (Neo) Training Hosting Deployment Frameworks Interfaces Infrastructure Amazon EC2 P3 & P3DN Amazon EC2 C5 FPGAs AWS Greengrass Amazon Elastic Inference Amazon Inferentia Labeling Model development Training Hosting ML SERVICES ML FRAMEWORKS & INFRASTRUCTURE New New New New New New New
  22. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Data Visualization & Analysis Business Problem – ML problem framing Data Collection Data Integration Data Preparation & Cleaning Feature Engineering Model Training & Parameter Tuning Model Evaluation Are Business Goals met? Model Deployment Monitoring & Debugging Yes No Data Augmentation Feature Augmentation The Machine Learning Process Re-training – Predictions
  23. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Data Visualization & Analysis Business Problem – ML problem framing Data Collection Data Integration Data Preparation & Cleaning Feature Engineering Model Training & Parameter Tuning Model Evaluation Are Business Goals met? Model Deployment Monitoring & Debugging Yes No Data Augmentation Feature Augmentation Integration: The Data Architecture Re-training • Build the data platform: • Amazon S3 • AWS Glue • Amazon Athena • Amazon EMR • Amazon Redshift Spectrum – Predictions
  24. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Data Visualization & Analysis Business Problem – ML problem framing Data Collection Data Integration Data Preparation & Cleaning Feature Engineering Model Training & Parameter Tuning Model Evaluation Are Business Goals met? Model Deployment Monitoring & Debugging Yes No Data Augmentation Feature Augmentation Why We built Amazon SageMaker: The Model Training Undifferentiated Heavy Lifting Re-training • Setup and manage Notebook Environments • Setup and manage Training Clusters • Write Data Connectors • Scale ML algorithms to large datasets • Distribute ML training algorithm to multiple machines • Secure Model artifacts – Predictions
  25. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker 1 2 3
  26. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark 1 2 3 Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
  27. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Successful models require high-quality data
  28. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark https://www.ebook5.net/blog/electroharmonix_fonts/
  29. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark https://www.ebook5.net/blog/electroharmonix_fonts/
  30. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark ٌؕٝծْؐٗծַٓٗشծٓ،ٌةװ https://www.ebook5.net/blog/electroharmonix_fonts/
  31. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Important things for Data Labeling • Good annotators • Easy-to-use tools • Clear communication with Annotators
  32. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
  33. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Model development options • Use the Amazon SageMaker built-in algorithm • Bring your own training script • Purchase a Machine Learning Model on AWS Marketplace
  34. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark - - - - /:: . : : . . /: /:: . / . :/ - : : / . - - - ) - - ( -
  35. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Model name Supervised? Algorithm description Descriptions Linear Learner Supervised Linear regression regression, classification XGBoost Supervised XGBoost (eXtreme Gradient Boosting) regression, classification PCA Unsupervised Principal Component Analysis Dimensions reduction k-means Unsupervised K-means Clustering k-NN Supervised K-nearest neighbour Clustering Factorization Machines Supervised Matrix factorization Recommend, regression, classification Random Cut Forest Unsupervised robust random cut tree Anomaly detection in time series LDA (Latent Dirichlet Allocation) Supervised Generative statistical model Topic modeling _.BDIJOFMFBSOJOHBMHPSJUINT_ SageMaker built-in algorithms ˟ -%"ךؔٔآشٕכ侄䌌ז׃ https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html
  36. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Field Model Supervised algorithms descriptions Image Processing Image classification supervised ResNet Image classification Object Detection supervised SSD (Single Shot multibox Detector) Detect multiple images with labeled bounding box Semantic Segmentation supervised FCN, PSP, DeepLabV3 (ResNet50, ResNet101) Detection of object region in image in pixel unit NLP seq2seq supervised Deep LSTM Text summaryvoice recognition Neural Topic Model unsupervised NTM, LDA Structuring text data Blazing text unsupervised Word2Vec Sentiment analysis supervised Text Classification Text mining Object2Vec supervised Genelized Word2Vec Classification, recommendation 5JNFTFSJFT DeepAR Forecasting supervised Autoregressive RNN Stochastic time series forecasting "OPNBMZ EFUFDUJPO IP Insights unsupervised NN Detection of malicious IP address _%FFQMFBSOJOHNPEFMT_ SageMaker build-in algorithms https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html
  37. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Bring your own ML algorithms Framework SageMaker container Deep learning TensorFlow Legacy mode: 1.4.1, 1.5.0, 1.6.0, 1.7.0, 1.8.0, 1.9.0, 1.10.0 Script mode: 1.11.0, 1.12.0 Chainer 4.0.0, 4.1.0, 5.0.0 PyTorch 0.4.0, 1.0.0 MXNet 1.3.0, 1.2.1, 1.1.0, 0.12.1 ML scikit-learn 0.20.0 https://github.com/aws/sagemaker-python-sdk/tree/master/src/sagemaker TensorFlow: https://github.com/aws/sagemaker-python-sdk/tree/master/src/sagemaker/tensorflow Chainer: https://github.com/aws/sagemaker-python-sdk/tree/master/src/sagemaker/chainer PyTorch: https://github.com/aws/sagemaker-python-sdk/tree/master/src/sagemaker/pytorch MXNet: https://github.com/aws/sagemaker-python-sdk/tree/master/src/sagemaker/mxnet Sklearn: https://github.com/aws/sagemaker-python-sdk/tree/master/src/sagemaker/sklearn • DL Frameworks supported in SageMaker
  38. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark AWS Marketplace for Machine Learning ML algorithms and models available instantly Subscribe in a single click Available in Amazon SageMaker K E Y F E A T U R E S Automatic labeling via machine learning IP protection Automated billing and metering Browse or search AWS Marketplace S E L L E R S Broad selection of paid, free, and open-source algorithms and models Data protection Discoverable on your AWS bill B U Y E R S
  39. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark • More than 200 algorithms for retail, media, etc. are available on AWS Marketplace for ML Algorithm Purchaser: • Training job and inference endpoint (also batch inference job ok) with Amazon SageMaker Algorithm Exhibitor: • Upload the model by concealing the contents of the model AWS Marketplace for Machine Learning
  40. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker customer use case Dely is running Japan's best cooking video service, Kurashiru. It strives every day to make culinary services that impact the world. Kurashiru helps many people per day, where it introduces a variety of tasty food recipes that color the dining table with cooking videos. Tens of millions of people watch and listen to the monthly recipe service in Japan. “We exceeded 15 million downloads of our mobile app, in 2.5 years since we launched the popular Kurashiru service. We believe it is critical to deliver the right content to our users at the right time using advanced technologies such as machine learning. To achieve this, we used Amazon SageMaker that helped us build and deploy the machine learning models in production in 90 days. We also improved the Click-Through Rate by 15% with content personalization”. - Masato Otake, CTO, Dely, Inc.
  41. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
  42. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
  43. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
  44. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Architecture of Amazon SageMaker AWS Cloud Office Network Notebook Instance Training instance Inference Inference Client SageMaker Service
  45. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Start Jupyter notebook for development environment AWS Cloud Office Network 1. Create Notebook instance from your local environment Notebook Instance Training instance Inference Inference 2. Store training script at local instance Client SageMaker Service
  46. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Launch Training Instance by docker container AWS Cloud Office Network 1. Run training job through SDK 4. Save trained models 2. Launch training instance 3. The training data and code are read on the container and the training is performed 5. When training is complete, instances are also automatically deleted Training instance Notebook Instance Client SageMaker Service
  47. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Endpoint hosting easily by docker container AWS Cloud Office Network 2. Launch inference instance 3. Loads model on container and acts as an endpoint SageMaker Service Hosting instance Notebook Instance Client 1. Create inference endpoint through SDK
  48. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark New machine learning capabilities in Amazon SageMaker to build, train and deploy with reinforcement learning
  49. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Reinforcement learning Achieve outcomes, not decisions Robotics Industrial controls Natural language systems Games
  50. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker Reinforcement Learning • 3-PO4BHF.BLFSJOUFHSBUFEXJUITJNVMBUJPOFOWJSPONFOUPGHBNJOHBOESPCPUJDT • 4VQQPSUT$PBDIBOE3-3BZBTSFJOGPSDFNFOUMFBSOJOHUPPMLJU • "843PCP.BLFS 3PCPTDIPPM FUDBWBJMBCMFWJB04"T0QFO"* (ZNJOUFSGBDF • %JTUSJCVUFEMFBSOJOHFOWJSPONFOU EJTUSJCVUJPOPGUSBJOJOHBOEFOWTJNVMBUJPO Container for Agent Container for Agent Container for environment Container for environment &OWJSPONFOU TJNVMBUPS RoboMaker, … OpenAI gym "HFOU RL tool kit Coach, RLLib Redis Action by policy Observation, reward Train policy
  51. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark SageMaker Reinforcement configuration ~ DeepRacer with RoboMaker ~  0CKFDUJWFGVODUJPO4FMGESJWJOHBDDPSEJOHUPMBOF DFOUFS  &OWJSPONFOU%ESJWJOHTJNVMBUPSIPTUFEBU"84 3PCP.BLFS  4UBUF107JNBHFTPGESJWFSTFOUUISPVHIDBNFSB PODBS  "DUJPOTEJGGFSFOUTUFFSJOHBOETQFFE DPOGJHVSBCMF  3FXBSE<FYBNQMF>QMVTXIFODPNFDMPTFSUP DFOUFSMBSHFQFOBMUZJNQPTFECZHFUUJOHPVUPGUIF USBDL1PTTJCMFUPTFUEFUBJMTTVDIBTBEEJOHB TUFFSJOHQFOBMUZ
  52. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark "84%FFQ3BDFS :PVDBOGJOEJO"NB[PODPN BU 4QFDJGJDBUJPO • TDBMFSBEJPDPOUSPMDBS • *OUFM"UPNQSPDFTTPS • .QJYFM QDBNFSB • 8J'J BD • #BUUFSZGPSDPNQVUF I BOE GPSNPUPS N • 6CVOUV-54 304 3PCPU 0QFSBUJOH4ZTUFN 0QFO7JOP https://aws.amazon.com/deepracer/
  53. © 2018, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Trademark Enhancing the fan experience One week of NFL games now creates 3TB of data. NFL uses Amazon SageMaker to analyze telemetry data to predict plays. Computations that could take months to refine now take only weeks or days. WATCH VIDEO >>
  54. © 2018, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Trademark Driving better healthcare outcomes Using Amazon SageMaker, GE Healthcare developed an ML model that can learn from thousands of medical scans to detect anomalies more accurately and efficiently, allowing radiologists to prioritize patients needing immediate attention.
  55. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
  56. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark Solve problems by Managed services Amazon SageMaker Environment • Building a development environment is troublesome (Development team launch uniform development environment, lack of labor) → Container • Difficult to estimate the necessary development and operation machine resources → Time charge · Auto scale • Need to easily create a CI / CD pipeline → with other AWS micro services Training PGNBDIJOFMFBSOJOHNPEFM • Build training model easily → Abundant SageMaker Examples • 4DBMBCMFSFTPVSDFTBOEJOUFSDPOOFDUJPOGPS%JTUSJCVUFE5SBJOJOH→ Easy distributed learning • .BOBHJOHFYQFSJNFOUTBOEEBUBNPEFMWFSTJPOJOH→ Simple data workflow, management tool Hosting &OEQPJOU • 1SFQBSJOHJOGFSFODFFOWJSPONFOUGPSNPEFMIPTUJOH→ Easy to deploy with Container • %JGGFSFOUTLJMMTFUUIBONBDIJOFMFBSOJOHJTSFRVJSFEUPEFQMPZJOQSPEVDUJPO → Container, Autoscale, Elastic Inference
  57. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark • Machine Learning to all developers • Benefits to run ML workload on AWS • Agility, Scalability, Cost optimization, effective pipeline with many services, Deploy worldwide instantly • Remove the undifferentiated heavy lifting • Use managed services to focus on differentiated valuable tasks • AWS Machine Learning stack • Various AI API Services • Use a fully managed environment with Amazon SageMaker
  58. © 2019, Amazon Web Services, Inc. or its Affiliates. All

    rights reserved. Amazon Confidential and Trademark