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

Google_Cloud_Next_19_AI_ML_Summary_public.pdf

soymsk
April 17, 2019

 Google_Cloud_Next_19_AI_ML_Summary_public.pdf

GCPNext'19 で発表された、AI/ML関連のまとめ資料 @社内勉強会

soymsk

April 17, 2019
Tweet

More Decks by soymsk

Other Decks in Technology

Transcript

  1. Google Cloud Next’19 Summary - AI/ML - Ryosuke Mashiko AI

    System Dept. Machine Learning Engineering Team DeNA Co,. Ltd
  2. Google Cloud Next’19 ▪ : San Francisco ▪ : 4/9

    ~ 4/11 ▪ GCP   ▪ : 3 4000 500
  3.  ▪   !, $ ✓ *)  ✓

    $& "# ▪ Uber/Lyft( ▪ +%'  FordBike Amazon Go
  4. tl;dr ▪ AI&1+0 3 (0  $% . ▪ Keyword:

    Digital Transformation( ‘18 Democratizing AI ) ✓ %'*4)/2: !%"%%etc ✓ $# 5- ▪ Anthos ✓ GKE both on GCP, ON-PREM, AWS ▪ Microsoft  ,6 ✓ Managed Active Directory ✓ Cloud SQL with Microsoft SQL server
  5. tl;dr ▪ AI/MLAI    ✓ AI Platform ✓

    AI Hub ✓ AutoML ◦ AutoML Table ◦ AutoML Video API ✓ BigQueryML
  6. BigQuery ML:   ~  ~ CREATE OR REPLACE

    MODEL `test_dataset.boston_predict_model` OPTIONS (model_type='linear_reg', input_label_cols=['price']) AS SELECT * FROM `test_dataset.boston_train` : http://soymsk.hatenablog.com/entry/2018/09/08/144846
  7. BigQuery ML:    ~  ~ SELECT *

    FROM ML.TRAINING_INFO(MODEL `test_dataset.boston_predict_model`)
  8. BigQuery ML:   ~ ~ SELECT * FROM ML.EVALUATE(MODEL

    `test_dataset.boston_predict_model`, ( SELECT * FROM `test_dataset.boston_test`))
  9. BigQuery ML:   ~ ~ SELECT price, predicted_price FROM

    ML.PREDICT(MODEL `soymsk-gcp.test_dataset.boston_predict_model`, ( SELECT * FROM `test_dataset.boston_test`))
  10. BigQuery DWH Preprocessing Model Train Predict BigQuery Data Model Train

    Predict Scheduled Query  BQML Scheduled Query  BQML ▪   BQ
  11. AutoML ▪ 20189Cloud Next FK +$ ▪ -6))+7 S ▪

    ,4+#(#0 >P%"*I5B; ✓ '++8GC"+C:  ✓ %"*GCP0 B;+! &QJ ✓ '++<OCAPI=/ ▪ APIE ✓ Vision … D13RA.?2 ✓ NL … I@LN ✓ Translation … HM
  12. NEW: AutoML Vision Edge ▪ AutoML    #"API$

    ▪ AutoML Vision Edge ✓ Vision   !   ◦ Tensorflow Lite format ◦ Edge TPU compiled TensorFlow Lite format ◦ TensorFlow ▪ Edge TPU     : https://cloud.google.com/edge-tpu/
  13. NEW: AutoML Video API ▪ Vision APIVideo API  ▪

     ✓  ✓   ✓    : https://cloud.google.com/vision/automl/docs/?hl=ja
  14. NEW: AutoML Table ▪ Structured Data(  ) API ✓

    BG/GCS ▪ AdaNet ✓ https://github.com/tensorflow/adanet  : https://cloud.google.com/automl-tables/?hl=ja
  15. AutoML Table:    ▪  1000  ▪

     > 1 hour  ▪   →
  16. AutoML Table:    ▪   ✓ IN:

    BQ/GIS ✓ OUT: BQ/GIS ▪    ✓ API
  17. AI Platform ▪ AI Platform (Cloud ML Engine $" +

    α) ▪ &0-# ✓  : Jupyter Notebook ( ≒ SageMaker @ AWS) ✓ AI Hub: %,+ !(  ✓ Kubeflow: *)%,  ✓ Training/Prediction:    %,/'1&/ ▪ CMLE*.
  18. AI Platform: Notebook ▪ Managed Jupyter (JaaS) ▪ DL 

    Tensorflow/Pytorch&! ✓ "$   ▪  $#&% ✓ GPU  
  19. AI Platform: AI Hub ▪ "  ! $) 

    "  ▪ Google+*#'  "  ( &%: https://aihub.cloud.google.com
  20. AI Platform: Kubeflow (Pipeline) ▪ *($/# 3.%1  OSS ✓

    TFX (Tensorflow Extended)  +- ✓ Kubernetes()  ✓ JupyterHub, Katib, Pipeline ▪ Kubeflow Pipeline ✓ ML# !,  )&  ✓ AI Hub2' ◦    AI Hub  "0 https://github.com/kubeflow/kubeflow
  21. TPU ▪ GoogleTensor Processing Unit ▪ GPU   ▪

    https://storage.googleapis.com/nexttpu/index.html ▪ Update: ✓ TPU v3 (2018)  GA
  22. ()" # TPU Multi GPU    AI Platform%

    DWH BigQuery Cloud Storage Cloud Machine Learning  # &   ! Kubeflow Pipeline EDA Model $ " # Single GPU  $ AI Notebook ( 4 0. 1 ) 4 22 0. ,   AI Hub
  23.  ▪ NEW: Recommendation API ✓ Personalized Recommendation API ◦

      ▪ NEW: Cloud Inference API ✓   β ✓   ◦ 
  24.  ▪ AI)80%/%AGPRQ ✓ @ IE$  ! ▪ Tensorflow')+,/7?!

    ✓ TF2.0 Eager Execution ModeKJNFMLD ✓ 5<-(1AC 4;>7=>(#SH !T ▪ 6:.(9&2 ✓ AIGCP*>3,AWS(9&2OB"  
  25.  : Cloud Dataflow ▪ Dataflow ✓ %*0 Apache Beam

    !2 ✓ Python/Java&MapReduce*!1 , ✓ %*4/(#()) !2 ▪ Update ✓ Preemptible VM(Spot instance@AWS) +3. ◦ 5" -$  ✓ ' 70%OFF