Driverless AI - Intro + Interactive Hands-on Lab
1 like710 views
The webinar by Arno Candel, CTO of H2O.ai, highlights the capabilities of Driverless AI in automating feature engineering and addressing the data scientist shortage. It showcases the platform's performance in various Kaggle competitions and outlines its features, including machine learning interpretation, automated visualization, and production deployment options. The roadmap for future enhancements, such as automated data cleaning and multi-GPU optimization, is also presented.
Driverless AI - Intro + Interactive Hands-on Lab
- 1. Driverless AI Webinar Arno Candel, PhD Chief Technology Officer H2O.ai, Inc. @arnocandel
- 4. Shortage of Data Scientists
- 5. Mistake Correction Automation needed to avoid human error
- 7. The “Secret Sauce” of Driverless AI: Feature Engineering https://www.youtube.com/watch?v=VMTKcT1iHww H2O.ai Webinar on Feature Engineering
- 8. Hours for Driverless AI — Weeks for grandmasters single run, fully automated: 6h on 3 GPUs Driverless AI: 18th place in private LB (out of 2926) Driverless AI: top 1% in BNP Paribas Kaggle competition Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 9. Driverless AI: top 5% in Amazon Kaggle competition Driverless AI: 80th place in private LB (out of 1687 - top 5%) With a little bit of stacking: 20th place (top 1.5%) Driverless AI produces feature engineering pipeline (“more columns”) for downstream use https://www.youtube.com/watch?v=qtUNyJlAID0&t=11s https://github.com/kaz-Anova/Competitive_Dai Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 10. Automatic Visualization Scalable outlier detection (no sampling) Contains novel statistical algorithms to only show “relevant” aspects of the data (coming soon: automated data cleaning) Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 11. Machine Learning Interpretation Gain confidence in models before deploying them! Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 12. MOJO: Pure Java Production Deployment • feature engineering and model scoring logic • auto-generated human-readable representation • minimal platform-independent storage format • scoring backend can be in any language (C/Java/C#/Go/etc.) Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 13. Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 14. Feature Now Q1 2018 Q2 2018 Q3 2018 AutoDL Feature Engineering Recipe Supervised Structured Data, CSV, Text Overfitting and Leakage Prevention Machine Learning Interpretation Automatic Visualization GUI Python client API Python scoring API HTTP Thrift Scoring API Multi-GPU (shared data) Scoring MOJO (100% Java or C) Data connectors: HDFS, SQL User Management: LDAP, Kerberos TensorFlow Deep Learning NLP Recipes Time Series Recipes Multi-GPU (sharded data) - optimized for DGX Volta UDR (User-Defined Recipes), Verticals Multi-Node Multi-GPU - optimized for DGX Volta Sparkling Water Backend for Driverless AI Driverless AI Roadmap Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 15. http://h2o.ai Webinars Copyright 2018 H2O.ai Inc. All Rights Reserved.
- 16. Hands-on Lab Copyright 2018 H2O.ai Inc. All Rights Reserved.