Machine Learning for everyone
0 likes623 views
This document discusses Amazon's use of machine learning across its businesses and services. It also discusses how Amazon uses machine learning to power product recommendations, customer support through self-service models, and optimizing call routing. Finally, it discusses how Amazon Machine Learning can help other companies build practical machine learning solutions to analyze data and gain business insights.
Machine Learning for everyone
- 1. © 2016, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Principal Technical Evangelist, AWS julsimon@amazon.fr @julsimon Julien Simon Machine Learning for everyone
- 5. Jeff Immelt, GE Chairman & CEO “If you went to bed last night as an industrial company, you’re going to wake up this morning as a software and analytics company.”
- 6. AMAZON.COM Building practical machine learning solutions
- 7. Machine Learning @ Amazon.com Amazon builds and uses Machine Learning solutions in hundreds of services across its various businesses. No off the shelf solution could handle the scale. Product recommendation is visible in many services.
- 8. Smart self-service Amazon also uses Machine Learning for Customer Support, building models based on recent orders, click- stream, user devices, prime membership usage, recent cases, recent account changes, etc. The models are used to provide efficient self-service to our customers. The self-service page is generated according to what your most likely question will be.
- 10. Smart customer call routing Not all customer interactions can be solved with self-service. Therefore, Amazon operates large support centers where Customer Service Representatives (CSR) handle customer requests. The Machine Learning models described above are used to optimize the human interactions of these requests. For example, they are used to route the customer call to the best CSR before the customer has even started to speak!
- 12. Well, we’re not Amazon. Can we still use this?
- 13. AMAZON WEB SERVICES Helping companies of all sizes build secure, scalable, and innovative applications
- 14. Let’s try to keep it simple Collect Store Analyze Consume Time to answer (Latency) Throughput Cost
- 16. Collect Store Analyze Consume A iOS Android Web Apps Logstash Amazon RDS Amazon DynamoDB Amazon ES Amazon S3 Apache Kafka Amazon Glacier Amazon Kinesis Amazon DynamoDB Amazon Redshift Impala Pig Amazon ML Streaming Amazon Kinesis AWS Lambda AmazonElasticMapReduce Amazon ElastiCache SearchSQLNoSQLCache StreamProcessing Batch Interactive Logging StreamStorage IoT Applications FileStorage Analysis&Visualization Hot Cold Warm Hot Slow Hot ML Fast Fast Amazon QuickSight Transactional Data File Data Stream Data Notebooks Predictions Apps & APIs Mobile Apps IDE Search Data ETL
- 17. Amazon Machine Learning Easy-to-use, managed machine learning service built for developers Robust, powerful machine learning technology based on Amazon’s internal systems Create models using your data already stored in the AWS Cloud Deploy models to production in seconds
- 18. BuildFax “Amazon Machine Learning democratizes the process of building predictive models. It's easy and fast to use, and has machine- learning best practices encapsulated in the product, which lets us deliver results significantly faster than in the past” Joe Emison, Founder & Chief Technology Officer https://aws.amazon.com/solutions/case-studies/buildfax/
- 19. Upserve Upserve is a software and mobile point of sale provider that offers a cloud-based restaurant management platform to restaurant owners across the U.S. “Using Amazon Machine Learning, we can predict the total number of customers who will walk through a restaurant’s doors in a night. As a result, restaurateurs can better prep and plan their staffing for that night” “It only took two weeks from the time we decided to use the technology to the moment we started using predictive data in the daily email we send out. And we immediately saw Amazon ML beating the baseline to predicting nightly covers” Bright Fulton, Director of Infrastructure Engineering https://aws.amazon.com/solutions/case-studies/upserve/
- 20. Alright, but we don’t want to build. Can we STILL use this?
- 21. PREDICSIS Machine Learning for everyone jeanlouis.fuccellaro@predicsis.com Jean-Louis Fuccellaro, CEO
- 22. Machine Learning for business Can I spot the right prospects ? What drives my upsell for this product ? Can I predict contract cancellations ?
- 23. Standard way StructureDataset Model/ Algorithm Preprocessing Model and validation Deployment Predictions -Feature engineering -Feature selection -Feature transformation -Missing value -Outlier handling -Variable type -Validate partioning -Choose model metric -Score models -Guess which algo to run -Model selection -Tuning hyper-parameters -Implementation in the application -Recoding in another language in production -Testing Other sources Importdata
- 24. Predicsis way BUSINESS INSIGHTS PREDICTIONS Yes Yes Yes … No Amazon S3 Amazon Redshift Other source
- 25. Out-band calls conversion rate. +30% ; 1 data analyst
- 26. Who to contact to prevent attrition ? +100% ; 1 business analyst
- 27. 1 / 1 / 10
- 28. Why should Big Data look like this… 2001, A Space Odyssey – © Turner Entertainment Company
- 29. When it could look like this? 2001, A Space Odyssey – © Turner Entertainment Company
- 30. Resources Big Data Whitepaper: http://bit.ly/2deGEVL Case studies: https://aws.amazon.com/solutions/case-studies/big-data/ Big Data Architectural Patterns and Best Practices on AWS https://www.youtube.com/watch?v=K7o5OlRLtvU Real-World Smart Applications With Amazon Machine Learning https://www.youtube.com/watch?v=sHJx1KJf8p0 Deep Learning: Going Beyond Machine Learning https://www.youtube.com/watch?v=Ra6m70d3t0o
- 31. AWS Enterprise Summit – 27/10/2016, Paris http://amzn.to/1X2yp0i
- 33. © 2016, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Thank you! Principal Technical Evangelist, AWS julsimon@amazon.fr @julsimon Julien Simon