Ce diaporama a bien été signalé.
Nous utilisons votre profil LinkedIn et vos données d’activité pour vous proposer des publicités personnalisées et pertinentes. Vous pouvez changer vos préférences de publicités à tout moment.

Solve complex business problems with Amazon Personalize and Amazon Forecast (April 2019)

343 vues

Publié le

Talk @ Builders Day / Dev Day 2019

Publié dans : Technologie
  • Soyez le premier à commenter

  • Soyez le premier à aimer ceci

Solve complex business problems with Amazon Personalize and Amazon Forecast (April 2019)

  1. 1. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Solve complex business problems with Amazon Forecast andAmazon Personalize Julien Simon Global Evangelist,AI & Machine Learning @julsimon
  2. 2. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  3. 3. Common applications& use cases Personalized recommendations Search reranking Notifications and emailsRelated Items
  4. 4. Personalizing user experience is proven to increase discoverability, engagement, user satisfaction, and revenue 30% of page views on Amazon are from recommendations … However, most customers find personalization hard to get right
  5. 5. Effective personalization requires solving multiple hard problems Reacting to user interactions in real time Avoiding mostly showing popular items Handling cold start (insufficient data about new users/items) Scale
  6. 6. Deep learning techniques have a direct impact on the bottom line SimilarityPopularity Neural network Matrix factorization +15.4% Engagement Recurrent Neural Net + Bandit Rule-based card ranker Bayesian network model +7.4% Engagement+29% Click Through +20% Click Through
  7. 7. DeepLearningdeliversstateof theartperformance 0.954 0.928 0.925 0.922 0.91 0.856 Rolling Average T-SVD [2009] PMF [2008] RRN [2017] DeepRec [2017] HRNN Ratings RMSE on Netflix 98 MM interactions, 500k users, 18k items Rolling Average T-SVD [2009] PMF [2008] RRN [2017] DeepRec [2017] HRNN 0.933 0.916 0.871 0.857 0.846 Rolling Average FM [2012] I-AutoRec [2015] RNN HRNN Ratings RMSE on MovieLens 20 MM interactions, 173k users, 131k items Rolling Average FM [2012] I-AutoRec [2015] RNN
  8. 8. AmazonPersonalize Improvecustomerexperienceswithpersonalizationandrecommendations K E Y F E A T U R E S Context-aware Recommendations Automated machine learning Continuous learning to improve performance
  9. 9. AmazonPersonalize:How itworks Amazon Personalize
  10. 10. FeedingdatatoAmazonPersonalize Historical user activity User attributes Item catalog Real-time data Mobile SDKs (coming soon) JavaScript SDK Amazon S3 bucket Server-Side SDKs Offline data
  11. 11. Traincustommodelsonceyouingestdata Use AutoML or pick a predefined algorithm recipes AutoML Hyper Parameter Optimization
  12. 12. Demo
  13. 13. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  14. 14. Sample use cases Product demand Workforce demand Financial metrics Inventory planning
  15. 15. Accuracy is the most important factor in forecasting Under-forecasting leads to lost opportunity Over-forecasting leads to wasted resources
  16. 16. Traditionaltime-seriesmodels
  17. 17. Traditionaltime-seriesmodels Trend
  18. 18. Traditionaltime-seriesmodels Trend + Seasonality
  19. 19. Traditionaltime-seriesmodels • Independent forecasts • Strong structural assumptions • De-facto industry standard • Well-understood, > 50 yrs. research • Data must match the structural assumptions • Cannot identify patterns across time series
  20. 20. Example Data Seasonality Trend Noise ? or useful information?
  21. 21. Traditional methods struggle with real-world forecasting Can’t handle time-series with no history Only process a single time- series at a time Don’t consider additional inputs: related time-series, metadata Only predict a single value: how trustworthy is it?
  22. 22. Can we do better?
  23. 23. 1–Multipletime-serieshelpidentifycommonpatterns
  24. 24. 2–Real-worldtimeseriesarenotwell-behaved…
  25. 25. …butusingadditionalinputshelpstofigurethemout
  26. 26. Usingadditionalinputs • Additional inputs can • Explain historical data • Drive forecast behavior • Examples from retail • Price information • Information about promotions • Out-of-stock information • Web page views • Known future events • Categorical inputs can be used to identify group-level patterns Fashion Women’s Clothing Shoes Watches Men’s Clothing Shoes Watches Girls' Clothing Shoes Watches Boys' Clothing Shoes Watches
  27. 27. 3–Thefuturecouldlooklikethis…
  28. 28. Or likethis..So howconfidentarewe?
  29. 29. Probabilisticforecasts:intervalsandconfidence
  30. 30. Probabilisticforecasts • Quantification of uncertainty • Support optimal decision making • Make “wrong” forecasts useful • Forecasts can be obtained for different quantiles of the predictive distribution p10: 10% of predictions with be lower p50: the mean value p90: 90% of predictions with be lower p10-p90 interval: 80% of possible predictions.
  31. 31. Deeplearningtime-seriesmodels • Global models: identify patterns using all available time series • Group-dependent seasonality and lifecycle • Behavior in response to extra inputs • Weak structural assumptions • Can be significantly more accurate than traditional methods • Can easily incorporate and learn from rich metadata • Support cold-start forecasts for new items
  32. 32. Usingdeeplearningincreasesforecastaccuracy MQ-RNN cold start spikes
  33. 33. AmazonForecast Improveforecastingaccuracybyupto50%at1/10ththecost K E Y F E A T U R E S Consider multiple time-series at once Automatic machine learning Visualize forecasts & import results into business apps Evaluate model accuracy Schedule forecasts and model retraining
  34. 34. Pre-definedschemasfordifferentbusinessdomains domains schemas
  35. 35. AmazonForecast:How itworks Target time-series Item meta-data tegory, genre, brand, etc. Amazon Forecast Related time-series
  36. 36. Traincustommodelsonceyouingestdata Use AutoML or pick a predefined algorithm recipes AutoML Hyper Parameter Optimization
  37. 37. KeymetricsreportedbyAmazonForecast
  38. 38. Demo https://github.com/aws-samples/amazon-forecast-samples
  39. 39. Getting started https://ml.aws https://aws.training/machinelearning https://aws.amazon.com/personalize https://aws.amazon.com/blogs/aws/amazon-personalize-real-time- personalization-and-recommendation-for-everyone/ https://aws.amazon.com/forecast https://aws.amazon.com/blogs/aws/amazon-forecast-time-series-forecasting- made-easy/
  40. 40. Thank you! © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Julien Simon Global Evangelist, AI & Machine Learning @julsimon https://medium.com/@julsimon

×